Compare commits

..

86 Commits
v0.3 ... v0.6.0

Author SHA1 Message Date
Thomas Roehl
31a38bc17d Update release action 2022-06-09 14:36:25 +02:00
Thomas Roehl
dbdec1eab8 Merge branch 'main' of github.com:ClusterCockpit/cc-metric-collector into main 2022-06-09 12:46:47 +02:00
Thomas Gruber
0d31ec481b Update Release.yml 2022-06-09 12:42:11 +02:00
Thomas Roehl
e22c3287e9 Merge branch 'main' of github.com:ClusterCockpit/cc-metric-collector into main 2022-06-08 15:26:05 +02:00
Thomas Gruber
8d85bd53f1 Merge latest development changes to main branch (#79)
* Cleanup: Remove unused code

* Use Golang duration parser for 'interval' and 'duration'
 in main config

* Update handling of LIKWID headers. Download only if not already present in the system. Fixes #73

* Units with cc-units (#64)

* Add option to normalize units with cc-unit

* Add unit conversion to router

* Add option to change unit prefix in the router

* Add to MetricRouter README

* Add order of operations in router to README

* Use second add_tags/del_tags only if metric gets renamed

* Skip disks in DiskstatCollector that have size=0

* Check readability of sensor files in TempCollector

* Fix for --once option

* Rename `cpu` type to `hwthread` (#69)

* Rename 'cpu' type to 'hwthread' to avoid naming clashes with MetricStore and CC-Webfrontend

* Collectors in parallel (#74)

* Provide info to CollectorManager whether the collector can be executed in parallel with others

* Split serial and parallel collectors. Read in parallel first

* Update NvidiaCollector with new metrics, MIG and NvLink support (#75)

* CC topology module update (#76)

* Rename CPU to hardware thread, write some comments

* Do renaming in other parts

* Remove CpuList and SocketList function from metricCollector. Available in ccTopology

* Option to use MIG UUID as subtype-id in NvidiaCollector

* Option to use MIG slice name as subtype-id in NvidiaCollector

* MetricRouter: Fix JSON in README

* Fix for Github Action to really use the selected version

* Remove Ganglia installation in runonce Action and add Go 1.18

* Fix daemon options in init script

* Add separate go.mod files to use it with deprecated 1.16

* Minor updates for Makefiles

* fix string comparison

* AMD ROCm SMI collector (#77)

* Add collector for AMD ROCm SMI metrics

* Fix import path

* Fix imports

* Remove Board Number

* store GPU index explicitly

* Remove board number from description

* Use http instead of ftp to download likwid

* Fix serial number in rocmCollector

* Improved http sink (#78)

* automatic flush in NatsSink

* tweak default options of HttpSink

* shorter cirt. section and retries for HttpSink

* fix error handling

* Remove file added by mistake.

* Use http instead of ftp to download likwid

* Fix serial number in rocmCollector

Co-authored-by: Thomas Roehl <thomas.roehl@fau.de>

Co-authored-by: Holger Obermaier <40787752+ho-ob@users.noreply.github.com>
Co-authored-by: Lou <lou.knauer@gmx.de>
2022-06-08 15:25:40 +02:00
Holger Obermaier
186a62a86b Fix: influx sink ignores config batch_size.
Feature: Add Redfish receiver
2022-05-04 13:00:56 +02:00
Holger Obermaier
e098c33179 Add some golang debug options 2022-05-04 12:48:46 +02:00
Thomas Roehl
38d4e0a730 Merge branch 'develop' of github.com:ClusterCockpit/cc-metric-collector into develop 2022-05-04 11:54:55 +02:00
Thomas Roehl
54d14519ca Skip mount points in DiskstatCollector if statfs() call does not work (bind mounts, ...) 2022-05-04 11:54:34 +02:00
Holger Obermaier
c35ac9dba8 Flush if batch size is reached 2022-05-04 11:28:06 +02:00
Holger Obermaier
c019f8e7ad Reuse tags and meta data tags 2022-05-03 17:55:33 +02:00
Holger Obermaier
fb6f6a4daa Fix GPFS collector last state handling 2022-05-02 16:57:19 +02:00
Holger Obermaier
9d6d0dbd93 Delete empty tags and meta data tags 2022-04-20 14:39:26 +02:00
Holger Obermaier
c2d4272fdf Clear workerInput channel after done event 2022-04-20 12:36:45 +02:00
Holger Obermaier
8c73095548 Allow to shutdown redfish receiver during metric read 2022-04-20 09:58:02 +02:00
Holger Obermaier
31c5c89a5a Fix: Close done channel 2022-04-19 14:01:23 +02:00
Holger Obermaier
bf9c7e1830 Update requirements 2022-04-19 12:15:51 +02:00
Holger Obermaier
48d34bf564 Adopt sinks.json for new meta_as_tags usage 2022-04-19 12:06:53 +02:00
Holger Obermaier
a1d85fa886 Add redfish receiver 2022-04-19 12:05:03 +02:00
Holger Obermaier
96ee16398e Removed unused done channel and wg wait group 2022-04-19 11:53:11 +02:00
Holger Obermaier
e7b8088c41 Extended go routine use case in sample receiver 2022-04-19 11:42:46 +02:00
Thomas Roehl
017cd58247 Updating page for LikwidCollector 2022-04-05 10:57:09 +02:00
Thomas Roehl
36dd440864 Merge branch 'develop' into main 2022-04-04 15:16:33 +02:00
Thomas Roehl
7b098e0b1b Fix for missing metrics in LikwidCollector is hwthread is inactive 2022-04-04 15:16:11 +02:00
Thomas Roehl
229a57b16a Merge branch 'develop' into main 2022-04-04 11:49:33 +02:00
Thomas Roehl
70a9530aba Set WriteFailedCallback to get some error message 2022-04-04 11:48:54 +02:00
Thomas Roehl
2f0b6057ca Merge branch 'develop' into main
- MetricRouter: Fix interval_timestamp option
- InfluxSink & InfluxAsyncSink: Add own flush mechanism
- InfluxSink & InfluxAsyncSink: Use default options, overwrite if
configured otherwise
- InfinibandCollector: Add units to metrics
- LikwidCollector: Fix for dying accessdaemon
2022-04-04 02:59:42 +02:00
Thomas Roehl
69f7c19659 InfluxAsyncSink: Add custom flush mechanism 2022-04-04 02:56:23 +02:00
Thomas Roehl
ecdb4c1bcf Add debug message when updating interval_timestep 2022-04-04 02:55:44 +02:00
Thomas Roehl
4d5b1adbc8 Fix for interval_timestamp option 2022-04-04 02:26:04 +02:00
Thomas Roehl
28348bd108 InfluxSink: Use batch&flush logic from HttpSink 2022-04-01 18:37:45 +02:00
Thomas Roehl
a3b9d8a90b HttpSink: Use sink name in error outputs 2022-04-01 18:36:54 +02:00
Thomas Roehl
7e43e9171e Use default options. Overwrite if anything is configured differently. Use seconds as precision 2022-04-01 17:26:56 +02:00
Thomas Roehl
5d25a7bf12 Add units to InfiniBandCollector 2022-04-01 17:14:26 +02:00
Thomas Roehl
83b4343310 Likwid receives signal at first Read, check when re-initializing 2022-04-01 17:10:31 +02:00
Thomas Roehl
f1d3cabdc6 Merge branch 'develop' of github.com:ClusterCockpit/cc-metric-collector into develop 2022-04-01 12:45:25 +02:00
Thomas Röhl
4763733d8d Merge branch 'develop' into main 2022-03-31 11:57:19 +02:00
Thomas Gruber
2a014b6fba Read unit of values from /proc/meminfo (#68) 2022-03-31 11:56:31 +02:00
Thomas Röhl
16e898ecca Merge branch 'develop' into main 2022-03-31 11:47:02 +02:00
Thomas Roehl
50479f9325 Move all LIKWID related stuff to late initialization routine 2022-03-24 18:12:23 +01:00
Thomas Roehl
e0e91844bc Use late initialization of LIKWID and catch access daemon death. Fixes #70 and fixes #71. 2022-03-24 17:56:51 +01:00
Thomas Roehl
296225f3a8 Always export all metrics in NfsCollectors 2022-03-24 13:50:35 +01:00
Thomas Roehl
43bcce6fb5 Merge branch 'develop' of github.com:ClusterCockpit/cc-metric-collector into develop 2022-03-22 18:05:38 +01:00
Thomas Roehl
622e94ae0e Fix DieList() if system does not support dies. Explicitly set entries in CpuData list 2022-03-22 15:58:10 +01:00
Thomas Roehl
c506114480 Add processing order to MetricRouter README and add missing options 2022-03-18 12:29:00 +01:00
Thomas Roehl
657543dded Ensure max_forward is at least 1 2022-03-18 12:28:52 +01:00
Thomas Roehl
4851382ad7 Merge branch 'develop' into main 2022-03-16 19:08:13 +01:00
Thomas Roehl
b66fdd1436 Add missing socket->thread_id map for LikwidCollector 2022-03-16 19:04:39 +01:00
Thomas Gruber
3f76947f54 Merge latest developments into main (#67)
* Update configuration.md

Add an additional receiver to have better alignment of components

* Change default GpfsCollector command to `mmpmon` (#53)

* Set default cmd to 'mmpmon'

* Reuse looked up path

* Cast const to string

* Just download LIKWID to get the headers (#54)

* Just download LIKWID to get the headers

* Remove perl-Data-Dumper from BuildRequires, only required by LIKWID build

* Add HttpReceiver as counterpart to the HttpSink (#49)

* Use GBytes as unit for large memory numbers

* Make maxForward configurable, save old name in meta in rename metrics and make the hostname tag key configurable

* Single release action (#55)

Building all RPMs and releasing in a single workflow

* Makefile target to build binary-only Debian packages (#61)

* Add 'install' and 'DEB' make targets to build binary-only Debian packages

* Add control file for DEB builds

* Use a single line for bash loop in make clean

* Add config options for retry intervals of InfluxDB clients (#59)

* Refactoring of LikwidCollector and metric units (#62)

* Reduce complexity of LikwidCollector and allow metric units

* Add unit to LikwidCollector docu and fix some typos

* Make library path configurable

* Use old metric name in Ganglia if rename has happened in the router (#60)

* Use old metric name if rename has happened in the router

* Also check for Ganglia renames for the oldname

* Derived metrics (#57)

* Add time-based derivatived (e.g. bandwidth) to some collectors

* Add documentation

* Add comments

* Fix: Only compute rates with a valid previous state

* Only compute rates with a valid previous state

* Define const values for net/dev fields

* Set default config values

* Add comments

* Refactor: Consolidate data structures

* Refactor: Consolidate data structures

* Refactor: Avoid struct deep copy

* Refactor: Avoid redundant tag maps

* Refactor: Use int64 type for absolut values

Co-authored-by: Holger Obermaier <40787752+ho-ob@users.noreply.github.com>

* Simplified iota usage

* Move unit tag to meta data tags

* Derived metrics (#65)

* Add time-based derivatived (e.g. bandwidth) to some collectors

* Add documentation

* Add comments

* Fix: Only compute rates with a valid previous state

* Only compute rates with a valid previous state

* Define const values for net/dev fields

* Set default config values

* Add comments

* Refactor: Consolidate data structures

* Refactor: Consolidate data structures

* Refactor: Avoid struct deep copy

* Refactor: Avoid redundant tag maps

* Refactor: Use int64 type for absolut values

* Update LustreCollector

Co-authored-by: Holger Obermaier <40787752+ho-ob@users.noreply.github.com>

* Meta to tags list and map for sinks (#63)

* Change ccMetric->Influx functions

* Use a meta_as_tags string list in config but create a lookup map afterwards

* Add meta as tag logic to sampleSink

* Fix staticcheck warnings (#66)

Co-authored-by: Holger Obermaier <40787752+ho-ob@users.noreply.github.com>
2022-03-15 16:41:11 +01:00
Thomas Gruber
c182d295f4 Fix staticcheck warnings (#66) 2022-03-15 16:38:20 +01:00
Thomas Roehl
beebcd7145 Merge branch 'develop' of github.com:ClusterCockpit/cc-metric-collector into develop 2022-03-15 16:16:39 +01:00
Thomas Gruber
57629a2e0a Meta to tags list and map for sinks (#63)
* Change ccMetric->Influx functions

* Use a meta_as_tags string list in config but create a lookup map afterwards

* Add meta as tag logic to sampleSink
2022-03-15 16:16:26 +01:00
Thomas Roehl
082eea525a Merge branch 'develop' of github.com:ClusterCockpit/cc-metric-collector into develop 2022-03-15 16:10:41 +01:00
Thomas Gruber
aa1afd745e Derived metrics (#65)
* Add time-based derivatived (e.g. bandwidth) to some collectors

* Add documentation

* Add comments

* Fix: Only compute rates with a valid previous state

* Only compute rates with a valid previous state

* Define const values for net/dev fields

* Set default config values

* Add comments

* Refactor: Consolidate data structures

* Refactor: Consolidate data structures

* Refactor: Avoid struct deep copy

* Refactor: Avoid redundant tag maps

* Refactor: Use int64 type for absolut values

* Update LustreCollector

Co-authored-by: Holger Obermaier <40787752+ho-ob@users.noreply.github.com>
2022-03-15 16:09:47 +01:00
Thomas Roehl
2b8266d1d2 Merge branch 'develop' of github.com:ClusterCockpit/cc-metric-collector into develop 2022-03-15 15:24:29 +01:00
Holger Obermaier
992b19d354 Move unit tag to meta data tags 2022-03-11 14:47:18 +01:00
Holger Obermaier
0b08ca9ae0 Simplified iota usage 2022-03-11 14:09:22 +01:00
Thomas Gruber
f6dae7c013 Derived metrics (#57)
* Add time-based derivatived (e.g. bandwidth) to some collectors

* Add documentation

* Add comments

* Fix: Only compute rates with a valid previous state

* Only compute rates with a valid previous state

* Define const values for net/dev fields

* Set default config values

* Add comments

* Refactor: Consolidate data structures

* Refactor: Consolidate data structures

* Refactor: Avoid struct deep copy

* Refactor: Avoid redundant tag maps

* Refactor: Use int64 type for absolut values

Co-authored-by: Holger Obermaier <40787752+ho-ob@users.noreply.github.com>
2022-03-11 13:48:18 +01:00
Thomas Gruber
1de3dda7be Use old metric name in Ganglia if rename has happened in the router (#60)
* Use old metric name if rename has happened in the router

* Also check for Ganglia renames for the oldname
2022-03-11 13:44:32 +01:00
Thomas Gruber
73f22c1041 Refactoring of LikwidCollector and metric units (#62)
* Reduce complexity of LikwidCollector and allow metric units

* Add unit to LikwidCollector docu and fix some typos

* Make library path configurable
2022-03-11 13:43:17 +01:00
Thomas Gruber
c9b8fcdaa7 Add config options for retry intervals of InfluxDB clients (#59) 2022-03-11 13:43:03 +01:00
Thomas Roehl
17f37583fc Use a single line for bash loop in make clean 2022-03-11 13:41:13 +01:00
Thomas Gruber
fb9ea992ea Makefile target to build binary-only Debian packages (#61)
* Add 'install' and 'DEB' make targets to build binary-only Debian packages

* Add control file for DEB builds
2022-03-11 13:39:09 +01:00
Thomas Gruber
21edca5f88 Single release action (#55)
Building all RPMs and releasing in a single workflow
2022-03-11 13:37:47 +01:00
Thomas Roehl
d835724d93 Merge branch 'develop' of github.com:ClusterCockpit/cc-metric-collector into develop 2022-03-09 17:02:58 +01:00
Thomas Roehl
3cf2f69a07 Make maxForward configurable, save old name in meta in rename metrics and make the hostname tag key configurable 2022-03-09 11:23:17 +01:00
Thomas Roehl
e7f7e68095 Use GBytes as unit for large memory numbers 2022-03-09 11:05:26 +01:00
Thomas Roehl
c5082bbffe Merge branch 'develop' of github.com:ClusterCockpit/cc-metric-collector into develop 2022-03-07 16:13:08 +01:00
Thomas Gruber
c0e600269a Add HttpReceiver as counterpart to the HttpSink (#49) 2022-03-05 17:30:55 +01:00
Thomas Gruber
f2486abeab Just download LIKWID to get the headers (#54)
* Just download LIKWID to get the headers

* Remove perl-Data-Dumper from BuildRequires, only required by LIKWID build
2022-03-05 17:30:40 +01:00
Thomas Gruber
21864e0ac4 Change default GpfsCollector command to mmpmon (#53)
* Set default cmd to 'mmpmon'

* Reuse looked up path

* Cast const to string
2022-03-05 14:42:04 +01:00
Thomas Roehl
3157386b3e Merge branch 'develop' into main 2022-03-04 23:34:28 +01:00
Thomas Gruber
8c668fcc6f Update configuration.md
Add an additional receiver to have better alignment of components
2022-03-04 18:33:57 +01:00
Thomas Roehl
1961edc659 Add documentation to help configuring the CC metric collector 2022-03-04 15:42:25 +01:00
Mehmet Soysal
547bc0461f Beegfs collector (#50)
* added beegfs collectors to collectors/README.md

* added beegfs collectors and docs

* added new beegfs collectors to AvailableCollectors list

* Feedback implemented

* changed error type

* changed error to only return

* changed beegfs lookup path

* fixed typo in md files

Co-authored-by: Mehmet Soysal <mehmet.soysal@kit.edu>
2022-03-04 14:35:47 +01:00
Thomas Roehl
7f62975a68 Set proper user for files 2022-03-04 11:52:48 +01:00
Thomas Roehl
5e64c01c9d Fix name for ClusterCockpit user 2022-03-04 11:51:21 +01:00
Thomas Roehl
f1d2828e1d Fix error print in LustreCollector 2022-03-04 11:32:10 +01:00
Thomas Roehl
07af660e50 Merge branch 'main' into develop 2022-03-03 17:33:13 +01:00
Thomas Roehl
af4c468ab5 Merge branch 'develop' of github.com:ClusterCockpit/cc-metric-collector into develop 2022-03-03 17:24:43 +01:00
Thomas Roehl
b3030a8d44 Use right systemd macro to create the user 2022-03-03 17:24:32 +01:00
Holger Obermaier
db04c8fbae Removed infinibandPerfQueryMetric.go. infinibandMetric.go offers the same functionality without requiring root privileges. 2022-03-03 15:52:50 +01:00
Thomas Roehl
4c1263312b Merge branch 'develop' of github.com:ClusterCockpit/cc-metric-collector into develop 2022-02-25 15:06:06 +01:00
Thomas Roehl
940623585c Merge branch 'develop' of github.com:ClusterCockpit/cc-metric-collector into develop 2022-02-25 13:53:34 +01:00
Thomas Roehl
87ecb12c6f Merge branch 'develop' of github.com:ClusterCockpit/cc-metric-collector into develop 2022-02-24 18:28:52 +01:00
Thomas Roehl
ae64eddcc8 Remove doubled import 2022-02-21 14:50:53 +01:00
84 changed files with 5103 additions and 1836 deletions

View File

@@ -3,6 +3,6 @@
"collectors" : ".github/ci-collectors.json",
"receivers" : ".github/ci-receivers.json",
"router" : ".github/ci-router.json",
"interval": 5,
"duration": 1
"interval": "5s",
"duration": "1s"
}

View File

@@ -1,6 +1,8 @@
{
"testoutput" : {
"type" : "stdout",
"meta_as_tags" : true
"meta_as_tags" : [
"unit"
]
}
}

View File

@@ -1,64 +0,0 @@
# See: https://docs.github.com/en/actions/using-workflows/workflow-syntax-for-github-actions
# Workflow name
name: AlmaLinux 8.5 RPM build
# Run on tag push
on:
push:
tags:
- '**'
jobs:
#
# Build on AlmaLinux 8.5 using go-toolset
#
AlmaLinux-RPM-build:
runs-on: ubuntu-latest
# See: https://hub.docker.com/_/almalinux
container: almalinux:8.5
steps:
# Use dnf to install development packages
- name: Install development packages
run: dnf --assumeyes group install "Development Tools" "RPM Development Tools"
# Checkout git repository and submodules
# fetch-depth must be 0 to use git describe
# See: https://github.com/marketplace/actions/checkout
- name: Checkout
uses: actions/checkout@v2
with:
submodules: recursive
fetch-depth: 0
# Use dnf to install build dependencies
- name: Install build dependencies
run: dnf --assumeyes builddep scripts/cc-metric-collector.spec
- name: RPM build MetricCollector
id: rpmbuild
run: make RPM
# See: https://github.com/actions/upload-artifact
- name: Save RPM as artifact
uses: actions/upload-artifact@v2
with:
name: cc-metric-collector RPM for AlmaLinux 8.5
path: ${{ steps.rpmbuild.outputs.RPM }}
- name: Save SRPM as artifact
uses: actions/upload-artifact@v2
with:
name: cc-metric-collector SRPM for AlmaLinux 8.5
path: ${{ steps.rpmbuild.outputs.SRPM }}
# See: https://github.com/softprops/action-gh-release
- name: Release
uses: softprops/action-gh-release@v1
if: startsWith(github.ref, 'refs/tags/')
with:
name: cc-metric-collector-${{github.ref_name}}
files: |
${{ steps.rpmbuild.outputs.RPM }}
${{ steps.rpmbuild.outputs.SRPM }}

View File

@@ -1,64 +0,0 @@
# See: https://docs.github.com/en/actions/using-workflows/workflow-syntax-for-github-actions
# Workflow name
name: Red Hat Universal Base Image 8 RPM build
# Run on tag push
on:
push:
tags:
- '**'
jobs:
#
# Build on UBI 8 using go-toolset
#
UBI-8-RPM-build:
runs-on: ubuntu-latest
# See: https://catalog.redhat.com/software/containers/ubi8/ubi/5c359854d70cc534b3a3784e?container-tabs=gti
container: registry.access.redhat.com/ubi8/ubi:8.5-226.1645809065
steps:
# Use dnf to install development packages
- name: Install development packages
run: dnf --assumeyes --disableplugin=subscription-manager install rpm-build go-srpm-macros rpm-build-libs rpm-libs gcc make python38 git
# Checkout git repository and submodules
# fetch-depth must be 0 to use git describe
# See: https://github.com/marketplace/actions/checkout
- name: Checkout
uses: actions/checkout@v2
with:
submodules: recursive
fetch-depth: 0
# Use dnf to install build dependencies
- name: Install build dependencies
run: dnf --assumeyes --disableplugin=subscription-manager builddep scripts/cc-metric-collector.spec
- name: RPM build MetricCollector
id: rpmbuild
run: make RPM
# See: https://github.com/actions/upload-artifact
- name: Save RPM as artifact
uses: actions/upload-artifact@v2
with:
name: cc-metric-collector RPM for UBI 8
path: ${{ steps.rpmbuild.outputs.RPM }}
- name: Save SRPM as artifact
uses: actions/upload-artifact@v2
with:
name: cc-metric-collector SRPM for UBI 8
path: ${{ steps.rpmbuild.outputs.SRPM }}
# See: https://github.com/softprops/action-gh-release
- name: Release
uses: softprops/action-gh-release@v1
if: startsWith(github.ref, 'refs/tags/')
with:
name: cc-metric-collector-${{github.ref_name}}
files: |
${{ steps.rpmbuild.outputs.RPM }}
${{ steps.rpmbuild.outputs.SRPM }}

187
.github/workflows/Release.yml vendored Normal file
View File

@@ -0,0 +1,187 @@
# See: https://docs.github.com/en/actions/using-workflows/workflow-syntax-for-github-actions
# Workflow name
name: Release
# Run on tag push
on:
push:
tags:
- '**'
jobs:
#
# Build on AlmaLinux 8.5 using go-toolset
#
AlmaLinux-RPM-build:
runs-on: ubuntu-latest
# See: https://hub.docker.com/_/almalinux
container: almalinux:8.5
# The job outputs link to the outputs of the 'rpmrename' step
# Only job outputs can be used in child jobs
outputs:
rpm : ${{steps.rpmrename.outputs.RPM}}
srpm : ${{steps.rpmrename.outputs.SRPM}}
steps:
# Use dnf to install development packages
- name: Install development packages
run: dnf --assumeyes group install "Development Tools" "RPM Development Tools"
# Checkout git repository and submodules
# fetch-depth must be 0 to use git describe
# See: https://github.com/marketplace/actions/checkout
- name: Checkout
uses: actions/checkout@v2
with:
submodules: recursive
fetch-depth: 0
# Use dnf to install build dependencies
- name: Install build dependencies
run: |
dnf --assumeyes install 'dnf-command(builddep)'
dnf --assumeyes install which
dnf --assumeyes builddep scripts/cc-metric-collector.spec
- name: RPM build MetricCollector
id: rpmbuild
run: make RPM
# AlmaLinux 8.5 is a derivate of RedHat Enterprise Linux 8 (UBI8),
# so the created RPM both contain the substring 'el8' in the RPM file names
# This step replaces the substring 'el8' to 'alma85'. It uses the move operation
# because it is unclear whether the default AlmaLinux 8.5 container contains the
# 'rename' command. This way we also get the new names for output.
- name: Rename RPMs (s/el8/alma85/)
id: rpmrename
run: |
OLD_RPM="${{steps.rpmbuild.outputs.RPM}}"
OLD_SRPM="${{steps.rpmbuild.outputs.SRPM}}"
NEW_RPM="${OLD_RPM/el8/alma85}"
NEW_SRPM=${OLD_SRPM/el8/alma85}
mv "${OLD_RPM}" "${NEW_RPM}"
mv "${OLD_SRPM}" "${NEW_SRPM}"
echo "::set-output name=SRPM::${NEW_SRPM}"
echo "::set-output name=RPM::${NEW_RPM}"
# See: https://github.com/actions/upload-artifact
- name: Save RPM as artifact
uses: actions/upload-artifact@v2
with:
name: cc-metric-collector RPM for AlmaLinux 8.5
path: ${{ steps.rpmrename.outputs.RPM }}
- name: Save SRPM as artifact
uses: actions/upload-artifact@v2
with:
name: cc-metric-collector SRPM for AlmaLinux 8.5
path: ${{ steps.rpmrename.outputs.SRPM }}
#
# Build on UBI 8 using go-toolset
#
UBI-8-RPM-build:
runs-on: ubuntu-latest
# See: https://catalog.redhat.com/software/containers/ubi8/ubi/5c359854d70cc534b3a3784e?container-tabs=gti
container: registry.access.redhat.com/ubi8/ubi:8.5-226.1645809065
# The job outputs link to the outputs of the 'rpmbuild' step
outputs:
rpm : ${{steps.rpmbuild.outputs.RPM}}
srpm : ${{steps.rpmbuild.outputs.SRPM}}
steps:
# Use dnf to install development packages
- name: Install development packages
run: dnf --assumeyes --disableplugin=subscription-manager install rpm-build go-srpm-macros rpm-build-libs rpm-libs gcc make python38 git
# Checkout git repository and submodules
# fetch-depth must be 0 to use git describe
# See: https://github.com/marketplace/actions/checkout
- name: Checkout
uses: actions/checkout@v2
with:
submodules: recursive
fetch-depth: 0
# Use dnf to install build dependencies
- name: Install build dependencies
run: dnf --assumeyes --disableplugin=subscription-manager builddep scripts/cc-metric-collector.spec
- name: RPM build MetricCollector
id: rpmbuild
run: make RPM
# See: https://github.com/actions/upload-artifact
- name: Save RPM as artifact
uses: actions/upload-artifact@v2
with:
name: cc-metric-collector RPM for UBI 8
path: ${{ steps.rpmbuild.outputs.RPM }}
- name: Save SRPM as artifact
uses: actions/upload-artifact@v2
with:
name: cc-metric-collector SRPM for UBI 8
path: ${{ steps.rpmbuild.outputs.SRPM }}
#
# Create release with fresh RPMs
#
Release:
runs-on: ubuntu-latest
# We need the RPMs, so add dependency
needs: [AlmaLinux-RPM-build, UBI-8-RPM-build]
steps:
# See: https://github.com/actions/download-artifact
- name: Download AlmaLinux 8.5 RPM
uses: actions/download-artifact@v2
with:
name: cc-metric-collector RPM for AlmaLinux 8.5
- name: Download AlmaLinux 8.5 SRPM
uses: actions/download-artifact@v2
with:
name: cc-metric-collector SRPM for AlmaLinux 8.5
- name: Download UBI 8 RPM
uses: actions/download-artifact@v2
with:
name: cc-metric-collector RPM for UBI 8
- name: Download UBI 8 SRPM
uses: actions/download-artifact@v2
with:
name: cc-metric-collector SRPM for UBI 8
# The download actions do not publish the name of the downloaded file,
# so we re-use the job outputs of the parent jobs. The files are all
# downloaded to the current folder.
# The gh-release action afterwards does not accept file lists but all
# files have to be listed at 'files'. The step creates one output per
# RPM package (2 per distro)
- name: Set RPM variables
id: files
run: |
ALMA_85_RPM=$(basename "${{ needs.AlmaLinux-RPM-build.outputs.rpm}}")
ALMA_85_SRPM=$(basename "${{ needs.AlmaLinux-RPM-build.outputs.srpm}}")
UBI_8_RPM=$(basename "${{ needs.UBI-8-RPM-build.outputs.rpm}}")
UBI_8_SRPM=$(basename "${{ needs.UBI-8-RPM-build.outputs.srpm}}")
echo "ALMA_85_RPM::${ALMA_85_RPM}"
echo "ALMA_85_SRPM::${ALMA_85_SRPM}"
echo "UBI_8_RPM::${UBI_8_RPM}"
echo "UBI_8_SRPM::${UBI_8_SRPM}"
echo "::set-output name=ALMA_85_RPM::${ALMA_85_RPM}"
echo "::set-output name=ALMA_85_SRPM::${ALMA_85_SRPM}"
echo "::set-output name=UBI_8_RPM::${UBI_8_RPM}"
echo "::set-output name=UBI_8_SRPM::${UBI_8_SRPM}"
# See: https://github.com/softprops/action-gh-release
- name: Release
uses: softprops/action-gh-release@v1
if: startsWith(github.ref, 'refs/tags/')
with:
name: cc-metric-collector-${{github.ref_name}}
files: |
${{ steps.files.outputs.ALMA_85_RPM }}
${{ steps.files.outputs.ALMA_85_SRPM }}
${{ steps.files.outputs.UBI_8_RPM }}
${{ steps.files.outputs.UBI_8_SRPM }}

View File

@@ -7,6 +7,32 @@ name: Run Test
on: push
jobs:
#
# Job build-1-18
# Build on latest Ubuntu using golang version 1.18
#
build-1-18:
runs-on: ubuntu-latest
steps:
# See: https://github.com/marketplace/actions/checkout
# Checkout git repository and submodules
- name: Checkout
uses: actions/checkout@v2
with:
submodules: recursive
# See: https://github.com/marketplace/actions/setup-go-environment
- name: Setup Golang
uses: actions/setup-go@v3
with:
go-version: '1.18.2'
- name: Build MetricCollector
run: make
- name: Run MetricCollector once
run: ./cc-metric-collector --once --config .github/ci-config.json
#
# Job build-1-17
# Build on latest Ubuntu using golang version 1.17
@@ -23,13 +49,9 @@ jobs:
# See: https://github.com/marketplace/actions/setup-go-environment
- name: Setup Golang
uses: actions/setup-go@v2
uses: actions/setup-go@v3
with:
go-version: '^1.17.7'
# Install libganglia
- name: Setup Ganglia
run: sudo apt install ganglia-monitor libganglia1
go-version: '1.17.7'
- name: Build MetricCollector
run: make
@@ -53,13 +75,9 @@ jobs:
# See: https://github.com/marketplace/actions/setup-go-environment
- name: Setup Golang
uses: actions/setup-go@v2
uses: actions/setup-go@v3
with:
go-version: '^1.16.7' # The version AlmaLinux 8.5 uses
# Install libganglia
- name: Setup Ganglia
run: sudo apt install ganglia-monitor libganglia1
go-version: '1.16.7' # The version AlmaLinux 8.5 uses
- name: Build MetricCollector
run: make

View File

@@ -15,47 +15,69 @@ COMPONENT_DIRS := collectors \
internal/ccTopology \
internal/multiChanTicker
BINDIR = bin
GOBIN = $(shell which go)
.PHONY: all
all: $(APP)
$(APP): $(GOSRC)
if [ "$(shell $(GOBIN) version | cut -d' ' -f 3 | cut -d '.' -f1-2)" = "go1.16" ]; then \
echo "1.16"; \
cp go.mod.1.16 go.mod; \
else \
echo "1.17+"; \
cp go.mod.1.17+ go.mod; \
fi
make -C collectors
go get
go build -o $(APP) $(GOSRC_APP)
$(GOBIN) get
$(GOBIN) build -o $(APP) $(GOSRC_APP)
install: $(APP)
@WORKSPACE=$(PREFIX)
@if [ -z "$${WORKSPACE}" ]; then exit 1; fi
@mkdir --parents --verbose $${WORKSPACE}/usr/$(BINDIR)
@install -Dpm 755 $(APP) $${WORKSPACE}/usr/$(BINDIR)/$(APP)
@mkdir --parents --verbose $${WORKSPACE}/etc/cc-metric-collector $${WORKSPACE}/etc/default $${WORKSPACE}/etc/systemd/system $${WORKSPACE}/etc/init.d
@install -Dpm 600 config.json $${WORKSPACE}/etc/cc-metric-collector/cc-metric-collector.json
@sed -i -e s+"\"./"+"\"/etc/cc-metric-collector/"+g $${WORKSPACE}/etc/cc-metric-collector/cc-metric-collector.json
@install -Dpm 600 sinks.json $${WORKSPACE}/etc/cc-metric-collector/sinks.json
@install -Dpm 600 collectors.json $${WORKSPACE}/etc/cc-metric-collector/collectors.json
@install -Dpm 600 router.json $${WORKSPACE}/etc/cc-metric-collector/router.json
@install -Dpm 600 receivers.json $${WORKSPACE}/etc/cc-metric-collector/receivers.json
@install -Dpm 600 scripts/cc-metric-collector.config $${WORKSPACE}/etc/default/cc-metric-collector
@install -Dpm 644 scripts/cc-metric-collector.service $${WORKSPACE}/etc/systemd/system/cc-metric-collector.service
@install -Dpm 644 scripts/cc-metric-collector.init $${WORKSPACE}/etc/init.d/cc-metric-collector
.PHONY: clean
.ONESHELL:
clean:
@for COMP in $(COMPONENT_DIRS)
do
if [[ -e $$COMP/Makefile ]]; then
make -C $$COMP clean
fi
done
@for COMP in $(COMPONENT_DIRS); do if [ -e $$COMP/Makefile ]; then make -C $$COMP clean; fi; done
rm -f $(APP)
.PHONY: fmt
fmt:
go fmt $(GOSRC_COLLECTORS)
go fmt $(GOSRC_SINKS)
go fmt $(GOSRC_RECEIVERS)
go fmt $(GOSRC_APP)
@for F in $(GOSRC_INTERNAL); do go fmt $$F; done
$(GOBIN) fmt $(GOSRC_COLLECTORS)
$(GOBIN) fmt $(GOSRC_SINKS)
$(GOBIN) fmt $(GOSRC_RECEIVERS)
$(GOBIN) fmt $(GOSRC_APP)
@for F in $(GOSRC_INTERNAL); do $(GOBIN) fmt $$F; done
# Examine Go source code and reports suspicious constructs
.PHONY: vet
vet:
go vet ./...
$(GOBIN) vet ./...
# Run linter for the Go programming language.
# Using static analysis, it finds bugs and performance issues, offers simplifications, and enforces style rules
.PHONY: staticcheck
staticcheck:
go install honnef.co/go/tools/cmd/staticcheck@latest
$$(go env GOPATH)/bin/staticcheck ./...
$(GOBIN) install honnef.co/go/tools/cmd/staticcheck@latest
$$($(GOBIN) env GOPATH)/bin/staticcheck ./...
.ONESHELL:
.PHONY: RPM
@@ -69,7 +91,7 @@ RPM: scripts/cc-metric-collector.spec
@COMMITISH="HEAD"
@VERS=$$(git describe --tags $${COMMITISH})
@VERS=$${VERS#v}
@VERS=$${VERS//-/_}
@VERS=$$(echo $$VERS | sed -e s+'-'+'_'+g)
@eval $$(rpmspec --query --queryformat "NAME='%{name}' VERSION='%{version}' RELEASE='%{release}' NVR='%{NVR}' NVRA='%{NVRA}'" --define="VERS $${VERS}" "$${SPECFILE}")
@PREFIX="$${NAME}-$${VERSION}"
@FORMAT="tar.gz"
@@ -86,3 +108,28 @@ RPM: scripts/cc-metric-collector.spec
@ echo "::set-output name=SRPM::$${SRPMFILE}"
@ echo "::set-output name=RPM::$${RPMFILE}"
@fi
.PHONY: DEB
DEB: scripts/cc-metric-collector.deb.control $(APP)
@BASEDIR=$${PWD}
@WORKSPACE=$${PWD}/.dpkgbuild
@DEBIANDIR=$${WORKSPACE}/debian
@DEBIANBINDIR=$${WORKSPACE}/DEBIAN
@mkdir --parents --verbose $$WORKSPACE $$DEBIANBINDIR
#@mkdir --parents --verbose $$DEBIANDIR
@CONTROLFILE="$${BASEDIR}/scripts/cc-metric-collector.deb.control"
@COMMITISH="HEAD"
@VERS=$$(git describe --tags --abbrev=0 $${COMMITISH})
@VERS=$${VERS#v}
@VERS=$$(echo $$VERS | sed -e s+'-'+'_'+g)
@ARCH=$$(uname -m)
@ARCH=$$(echo $$ARCH | sed -e s+'_'+'-'+g)
@PREFIX="$${NAME}-$${VERSION}_$${ARCH}"
@SIZE_BYTES=$$(du -bcs --exclude=.dpkgbuild "$$WORKSPACE"/ | awk '{print $$1}' | head -1 | sed -e 's/^0\+//')
@SIZE="$$(awk -v size="$$SIZE_BYTES" 'BEGIN {print (size/1024)+1}' | awk '{print int($$0)}')"
#@sed -e s+"{VERSION}"+"$$VERS"+g -e s+"{INSTALLED_SIZE}"+"$$SIZE"+g -e s+"{ARCH}"+"$$ARCH"+g $$CONTROLFILE > $${DEBIANDIR}/control
@sed -e s+"{VERSION}"+"$$VERS"+g -e s+"{INSTALLED_SIZE}"+"$$SIZE"+g -e s+"{ARCH}"+"$$ARCH"+g $$CONTROLFILE > $${DEBIANBINDIR}/control
@make PREFIX=$${WORKSPACE} install
@DEB_FILE="cc-metric-collector_$${VERS}_$${ARCH}.deb"
@dpkg-deb -b $${WORKSPACE} "$$DEB_FILE"
@rm -r "$${WORKSPACE}"

View File

@@ -22,8 +22,8 @@ import (
)
type CentralConfigFile struct {
Interval int `json:"interval"`
Duration int `json:"duration"`
Interval string `json:"interval"`
Duration string `json:"duration"`
CollectorConfigFile string `json:"collectors"`
RouterConfigFile string `json:"router"`
SinkConfigFile string `json:"sinks"`
@@ -173,16 +173,36 @@ func mainFunc() int {
cclog.Error("Error reading configuration file ", rcfg.CliArgs["configfile"], ": ", err.Error())
return 1
}
if rcfg.ConfigFile.Interval <= 0 || time.Duration(rcfg.ConfigFile.Interval)*time.Second <= 0 {
cclog.Error("Configuration value 'interval' must be greater than zero")
// Properly use duration parser with inputs like '60s', '5m' or similar
if len(rcfg.ConfigFile.Interval) > 0 {
t, err := time.ParseDuration(rcfg.ConfigFile.Interval)
if err != nil {
cclog.Error("Configuration value 'interval' no valid duration")
}
rcfg.Interval = t
if rcfg.Interval == 0 {
cclog.Error("Configuration value 'interval' must be greater than zero")
return 1
}
}
// Properly use duration parser with inputs like '60s', '5m' or similar
if len(rcfg.ConfigFile.Duration) > 0 {
t, err := time.ParseDuration(rcfg.ConfigFile.Duration)
if err != nil {
cclog.Error("Configuration value 'duration' no valid duration")
}
rcfg.Duration = t
if rcfg.Duration == 0 {
cclog.Error("Configuration value 'duration' must be greater than zero")
return 1
}
}
if rcfg.Duration > rcfg.Interval {
cclog.Error("The interval should be greater than duration")
return 1
}
rcfg.Interval = time.Duration(rcfg.ConfigFile.Interval) * time.Second
if rcfg.ConfigFile.Duration <= 0 || time.Duration(rcfg.ConfigFile.Duration)*time.Second <= 0 {
cclog.Error("Configuration value 'duration' must be greater than zero")
return 1
}
rcfg.Duration = time.Duration(rcfg.ConfigFile.Duration) * time.Second
if len(rcfg.ConfigFile.RouterConfigFile) == 0 {
cclog.Error("Metric router configuration file must be set")
@@ -271,7 +291,7 @@ func mainFunc() int {
// Wait until one tick has passed. This is a workaround
if rcfg.CliArgs["once"] == "true" {
x := 1.2 * float64(rcfg.ConfigFile.Interval)
x := 1.2 * float64(rcfg.Interval.Seconds())
time.Sleep(time.Duration(int(x)) * time.Second)
shutdownSignal <- os.Interrupt
}

View File

@@ -12,6 +12,12 @@
"proc_total"
]
},
"netstat": {
"include_devices": [
"enp5s0"
],
"send_derived_values": true
},
"numastats": {},
"nvidia": {},
"tempstat": {

View File

@@ -1,79 +1,31 @@
# Use central installation
CENTRAL_INSTALL = false
# How to access hardware performance counters through LIKWID.
# Recommended is 'direct' mode
ACCESSMODE = direct
#######################################################################
# if CENTRAL_INSTALL == true
#######################################################################
# Path to central installation (if CENTRAL_INSTALL=true)
LIKWID_BASE=/apps/likwid/5.2.1
# LIKWID version (should be same major version as central installation, 5.2.x)
# LIKWID version
LIKWID_VERSION = 5.2.1
LIKWID_INSTALLED_FOLDER=$(shell dirname $(shell which likwid-topology 2>/dev/null) 2>/dev/null)
#######################################################################
# if CENTRAL_INSTALL == false and ACCESSMODE == accessdaemon
#######################################################################
# Where to install the accessdaemon
DAEMON_INSTALLDIR = /usr/local
# Which user to use for the accessdaemon
DAEMON_USER = root
# Which group to use for the accessdaemon
DAEMON_GROUP = root
LIKWID_FOLDER="$(shell pwd)/likwid"
all: $(LIKWID_FOLDER)/likwid.h
.ONESHELL:
.PHONY: $(LIKWID_FOLDER)/likwid.h
$(LIKWID_FOLDER)/likwid.h:
if [ "$(LIKWID_INSTALLED_FOLDER)" != "" ]; then \
BASE="$(LIKWID_INSTALLED_FOLDER)/../include"; \
mkdir -p $(LIKWID_FOLDER); \
cp $$BASE/*.h $(LIKWID_FOLDER); \
else \
BUILD_FOLDER="$${PWD}/likwidbuild"; \
if [ -d $(LIKWID_FOLDER) ]; then rm -r $(LIKWID_FOLDER); fi; \
mkdir --parents --verbose $(LIKWID_FOLDER) $${BUILD_FOLDER}; \
wget -P "$${BUILD_FOLDER}" http://ftp.rrze.uni-erlangen.de/mirrors/likwid/likwid-$(LIKWID_VERSION).tar.gz; \
tar -C $${BUILD_FOLDER} -xf $${BUILD_FOLDER}/likwid-$(LIKWID_VERSION).tar.gz; \
install -Dpm 0644 $${BUILD_FOLDER}/likwid-$(LIKWID_VERSION)/src/includes/likwid*.h $(LIKWID_FOLDER)/; \
install -Dpm 0644 $${BUILD_FOLDER}/likwid-$(LIKWID_VERSION)/src/includes/bstrlib.h $(LIKWID_FOLDER)/; \
rm -r $${BUILD_FOLDER}; \
fi
#################################################
# No need to change anything below this line
#################################################
INSTALL_FOLDER = ./likwid
BUILD_FOLDER = ./likwid/build
ifneq ($(strip $(CENTRAL_INSTALL)),true)
LIKWID_BASE := $(shell pwd)/$(INSTALL_FOLDER)
DAEMON_BASE := $(LIKWID_BASE)
GROUPS_BASE := $(LIKWID_BASE)/groups
all: $(INSTALL_FOLDER)/liblikwid.a cleanup
else
DAEMON_BASE= $(LIKWID_BASE)/sbin
all: $(INSTALL_FOLDER)/liblikwid.a cleanup
endif
$(BUILD_FOLDER)/likwid-$(LIKWID_VERSION).tar.gz: $(BUILD_FOLDER)
wget -P $(BUILD_FOLDER) ftp://ftp.rrze.uni-erlangen.de/mirrors/likwid/likwid-$(LIKWID_VERSION).tar.gz
$(BUILD_FOLDER):
mkdir -p $(BUILD_FOLDER)
$(INSTALL_FOLDER):
mkdir -p $(INSTALL_FOLDER)
$(BUILD_FOLDER)/likwid-$(LIKWID_VERSION): $(BUILD_FOLDER)/likwid-$(LIKWID_VERSION).tar.gz
tar -C $(BUILD_FOLDER) -xf $(BUILD_FOLDER)/likwid-$(LIKWID_VERSION).tar.gz
$(INSTALL_FOLDER)/liblikwid.a: $(BUILD_FOLDER)/likwid-$(LIKWID_VERSION) $(INSTALL_FOLDER)
cd "$(BUILD_FOLDER)/likwid-$(LIKWID_VERSION)" && make "PREFIX=$(LIKWID_BASE)" "SHARED_LIBRARY=false" "ACCESSMODE=$(ACCESSMODE)" "INSTALLED_ACCESSDAEMON=$(DAEMON_INSTALLDIR)/likwid-accessD"
cp \
$(BUILD_FOLDER)/likwid-$(LIKWID_VERSION)/liblikwid.a \
$(BUILD_FOLDER)/likwid-$(LIKWID_VERSION)/ext/hwloc/liblikwid-hwloc.a \
$(BUILD_FOLDER)/likwid-$(LIKWID_VERSION)/src/includes/likwid*.h \
$(BUILD_FOLDER)/likwid-$(LIKWID_VERSION)/src/includes/bstrlib.h \
$(INSTALL_FOLDER)
$(DAEMON_INSTALLDIR)/likwid-accessD: $(BUILD_FOLDER)/likwid-$(LIKWID_VERSION)/likwid-accessD
sudo -u $(DAEMON_USER) -g $(DAEMON_GROUP) install -m 4775 $(BUILD_FOLDER)/likwid-$(LIKWID_VERSION)/likwid-accessD $(DAEMON_INSTALLDIR)/likwid-accessD
prepare_collector: likwidMetric.go
cp likwidMetric.go likwidMetric.go.orig
sed -i -e s+"const GROUPPATH =.*"+"const GROUPPATH = \`$(GROUPS_BASE)\`"+g likwidMetric.go
cleanup:
rm -rf $(BUILD_FOLDER)
clean: cleanup
clean:
rm -rf likwid
.PHONY: clean

View File

@@ -37,6 +37,9 @@ In contrast to the configuration files for sinks and receivers, the collectors c
* [`cpufreq_cpuinfo`](./cpufreqCpuinfoMetric.md)
* [`numastat`](./numastatMetric.md)
* [`gpfs`](./gpfsMetric.md)
* [`beegfs_meta`](./beegfsmetaMetric.md)
* [`beegfs_storage`](./beegfsstorageMetric.md)
* [`rocm_smi`](./rocmsmiMetric.md)
## Todos

View File

@@ -0,0 +1,230 @@
package collectors
import (
"bufio"
"bytes"
"encoding/json"
"fmt"
"io/ioutil"
"os"
"os/exec"
"os/user"
"regexp"
"strconv"
"strings"
"time"
cclog "github.com/ClusterCockpit/cc-metric-collector/internal/ccLogger"
lp "github.com/ClusterCockpit/cc-metric-collector/internal/ccMetric"
)
const DEFAULT_BEEGFS_CMD = "beegfs-ctl"
// Struct for the collector-specific JSON config
type BeegfsMetaCollectorConfig struct {
Beegfs string `json:"beegfs_path"`
ExcludeMetrics []string `json:"exclude_metrics,omitempty"`
ExcludeFilesystem []string `json:"exclude_filesystem"`
}
type BeegfsMetaCollector struct {
metricCollector
tags map[string]string
matches map[string]string
config BeegfsMetaCollectorConfig
skipFS map[string]struct{}
}
func (m *BeegfsMetaCollector) Init(config json.RawMessage) error {
// Check if already initialized
if m.init {
return nil
}
// Metrics
var nodeMdstat_array = [39]string{
"sum", "ack", "close", "entInf",
"fndOwn", "mkdir", "create", "rddir",
"refrEn", "mdsInf", "rmdir", "rmLnk",
"mvDirIns", "mvFiIns", "open", "ren",
"sChDrct", "sAttr", "sDirPat", "stat",
"statfs", "trunc", "symlnk", "unlnk",
"lookLI", "statLI", "revalLI", "openLI",
"createLI", "hardlnk", "flckAp", "flckEn",
"flckRg", "dirparent", "listXA", "getXA",
"rmXA", "setXA", "mirror"}
m.name = "BeegfsMetaCollector"
m.setup()
m.parallel = true
// Set default beegfs-ctl binary
m.config.Beegfs = DEFAULT_BEEGFS_CMD
// Read JSON configuration
if len(config) > 0 {
err := json.Unmarshal(config, &m.config)
if err != nil {
return err
}
}
//create map with possible variables
m.matches = make(map[string]string)
for _, value := range nodeMdstat_array {
_, skip := stringArrayContains(m.config.ExcludeMetrics, value)
if skip {
m.matches["other"] = "0"
} else {
m.matches["beegfs_cmeta_"+value] = "0"
}
}
m.meta = map[string]string{
"source": m.name,
"group": "BeegfsMeta",
}
m.tags = map[string]string{
"type": "node",
"filesystem": "",
}
m.skipFS = make(map[string]struct{})
for _, fs := range m.config.ExcludeFilesystem {
m.skipFS[fs] = struct{}{}
}
// Beegfs file system statistics can only be queried by user root
user, err := user.Current()
if err != nil {
return fmt.Errorf("BeegfsMetaCollector.Init(): Failed to get current user: %v", err)
}
if user.Uid != "0" {
return fmt.Errorf("BeegfsMetaCollector.Init(): BeeGFS file system statistics can only be queried by user root")
}
// Check if beegfs-ctl is in executable search path
_, err = exec.LookPath(m.config.Beegfs)
if err != nil {
return fmt.Errorf("BeegfsMetaCollector.Init(): Failed to find beegfs-ctl binary '%s': %v", m.config.Beegfs, err)
}
m.init = true
return nil
}
func (m *BeegfsMetaCollector) Read(interval time.Duration, output chan lp.CCMetric) {
if !m.init {
return
}
//get mounpoint
buffer, _ := ioutil.ReadFile(string("/proc/mounts"))
mounts := strings.Split(string(buffer), "\n")
var mountpoints []string
for _, line := range mounts {
if len(line) == 0 {
continue
}
f := strings.Fields(line)
if strings.Contains(f[0], "beegfs_ondemand") {
// Skip excluded filesystems
if _, skip := m.skipFS[f[1]]; skip {
continue
}
mountpoints = append(mountpoints, f[1])
}
}
if len(mountpoints) == 0 {
return
}
for _, mountpoint := range mountpoints {
m.tags["filesystem"] = mountpoint
// bwwgfs-ctl:
// --clientstats: Show client IO statistics.
// --nodetype=meta: The node type to query (meta, storage).
// --interval:
// --mount=/mnt/beeond/: Which mount point
//cmd := exec.Command(m.config.Beegfs, "/root/mc/test.txt")
mountoption := "--mount=" + mountpoint
cmd := exec.Command(m.config.Beegfs, "--clientstats",
"--nodetype=meta", mountoption, "--allstats")
cmd.Stdin = strings.NewReader("\n")
cmdStdout := new(bytes.Buffer)
cmdStderr := new(bytes.Buffer)
cmd.Stdout = cmdStdout
cmd.Stderr = cmdStderr
err := cmd.Run()
if err != nil {
fmt.Fprintf(os.Stderr, "BeegfsMetaCollector.Read(): Failed to execute command \"%s\": %s\n", cmd.String(), err.Error())
fmt.Fprintf(os.Stderr, "BeegfsMetaCollector.Read(): command exit code: \"%d\"\n", cmd.ProcessState.ExitCode())
data, _ := ioutil.ReadAll(cmdStderr)
fmt.Fprintf(os.Stderr, "BeegfsMetaCollector.Read(): command stderr: \"%s\"\n", string(data))
data, _ = ioutil.ReadAll(cmdStdout)
fmt.Fprintf(os.Stderr, "BeegfsMetaCollector.Read(): command stdout: \"%s\"\n", string(data))
return
}
// Read I/O statistics
scanner := bufio.NewScanner(cmdStdout)
sumLine := regexp.MustCompile(`^Sum:\s+\d+\s+\[[a-zA-Z]+\]+`)
//Line := regexp.MustCompile(`^(.*)\s+(\d)+\s+\[([a-zA-Z]+)\]+`)
statsLine := regexp.MustCompile(`^(.*?)\s+?(\d.*?)$`)
singleSpacePattern := regexp.MustCompile(`\s+`)
removePattern := regexp.MustCompile(`[\[|\]]`)
for scanner.Scan() {
readLine := scanner.Text()
//fmt.Println(readLine)
// Jump few lines, we only want the I/O stats from nodes
if !sumLine.MatchString(readLine) {
continue
}
match := statsLine.FindStringSubmatch(readLine)
// nodeName = "Sum:" or would be nodes
// nodeName := match[1]
//Remove multiple whitespaces
dummy := removePattern.ReplaceAllString(match[2], " ")
metaStats := strings.TrimSpace(singleSpacePattern.ReplaceAllString(dummy, " "))
split := strings.Split(metaStats, " ")
// fill map with values
// split[i+1] = mdname
// split[i] = amount of md operations
for i := 0; i <= len(split)-1; i += 2 {
if _, ok := m.matches[split[i+1]]; ok {
m.matches["beegfs_cmeta_"+split[i+1]] = split[i]
} else {
f1, err := strconv.ParseFloat(m.matches["other"], 32)
if err != nil {
cclog.ComponentError(
m.name,
fmt.Sprintf("Metric (other): Failed to convert str written '%s' to float: %v", m.matches["other"], err))
continue
}
f2, err := strconv.ParseFloat(split[i], 32)
if err != nil {
cclog.ComponentError(
m.name,
fmt.Sprintf("Metric (other): Failed to convert str written '%s' to float: %v", m.matches["other"], err))
continue
}
//mdStat["other"] = fmt.Sprintf("%f", f1+f2)
m.matches["beegfs_cstorage_other"] = fmt.Sprintf("%f", f1+f2)
}
}
for key, data := range m.matches {
value, _ := strconv.ParseFloat(data, 32)
y, err := lp.New(key, m.tags, m.meta, map[string]interface{}{"value": value}, time.Now())
if err == nil {
output <- y
}
}
}
}
}
func (m *BeegfsMetaCollector) Close() {
m.init = false
}

View File

@@ -0,0 +1,75 @@
## `BeeGFS on Demand` collector
This Collector is to collect BeeGFS on Demand (BeeOND) metadata clientstats.
```json
"beegfs_meta": {
"beegfs_path": "/usr/bin/beegfs-ctl",
"exclude_filesystem": [
"/mnt/ignore_me"
],
"exclude_metrics": [
"ack",
"entInf",
"fndOwn"
]
}
```
The `BeeGFS On Demand (BeeOND)` collector uses the `beegfs-ctl` command to read performance metrics for
BeeGFS filesystems.
The reported filesystems can be filtered with the `exclude_filesystem` option
in the configuration.
The path to the `beegfs-ctl` command can be configured with the `beegfs_path` option
in the configuration.
When using the `exclude_metrics` option, the excluded metrics are summed as `other`.
Important: The metrics listed below, are similar to the naming of BeeGFS. The Collector prefixes these with `beegfs_cstorage`(beegfs client storage).
For example beegfs metric `open`-> `beegfs_cstorage_open`
Available Metrics:
* sum
* ack
* close
* entInf
* fndOwn
* mkdir
* create
* rddir
* refrEnt
* mdsInf
* rmdir
* rmLnk
* mvDirIns
* mvFiIns
* open
* ren
* sChDrct
* sAttr
* sDirPat
* stat
* statfs
* trunc
* symlnk
* unlnk
* lookLI
* statLI
* revalLI
* openLI
* createLI
* hardlnk
* flckAp
* flckEn
* flckRg
* dirparent
* listXA
* getXA
* rmXA
* setXA
* mirror
The collector adds a `filesystem` tag to all metrics

View File

@@ -0,0 +1,222 @@
package collectors
import (
"bufio"
"bytes"
"encoding/json"
"fmt"
"io/ioutil"
"os"
"os/exec"
"os/user"
"regexp"
"strconv"
"strings"
"time"
cclog "github.com/ClusterCockpit/cc-metric-collector/internal/ccLogger"
lp "github.com/ClusterCockpit/cc-metric-collector/internal/ccMetric"
)
// Struct for the collector-specific JSON config
type BeegfsStorageCollectorConfig struct {
Beegfs string `json:"beegfs_path"`
ExcludeMetrics []string `json:"exclude_metrics,omitempty"`
ExcludeFilesystem []string `json:"exclude_filesystem"`
}
type BeegfsStorageCollector struct {
metricCollector
tags map[string]string
matches map[string]string
config BeegfsStorageCollectorConfig
skipFS map[string]struct{}
}
func (m *BeegfsStorageCollector) Init(config json.RawMessage) error {
// Check if already initialized
if m.init {
return nil
}
// Metrics
var storageStat_array = [18]string{
"sum", "ack", "sChDrct", "getFSize",
"sAttr", "statfs", "trunc", "close",
"fsync", "ops-rd", "MiB-rd/s", "ops-wr",
"MiB-wr/s", "gendbg", "hrtbeat", "remNode",
"storInf", "unlnk"}
m.name = "BeegfsStorageCollector"
m.setup()
m.parallel = true
// Set default beegfs-ctl binary
m.config.Beegfs = DEFAULT_BEEGFS_CMD
// Read JSON configuration
if len(config) > 0 {
err := json.Unmarshal(config, &m.config)
if err != nil {
return err
}
}
println(m.config.Beegfs)
//create map with possible variables
m.matches = make(map[string]string)
for _, value := range storageStat_array {
_, skip := stringArrayContains(m.config.ExcludeMetrics, value)
if skip {
m.matches["other"] = "0"
} else {
m.matches["beegfs_cstorage_"+value] = "0"
}
}
m.meta = map[string]string{
"source": m.name,
"group": "BeegfsStorage",
}
m.tags = map[string]string{
"type": "node",
"filesystem": "",
}
m.skipFS = make(map[string]struct{})
for _, fs := range m.config.ExcludeFilesystem {
m.skipFS[fs] = struct{}{}
}
// Beegfs file system statistics can only be queried by user root
user, err := user.Current()
if err != nil {
return fmt.Errorf("BeegfsStorageCollector.Init(): Failed to get current user: %v", err)
}
if user.Uid != "0" {
return fmt.Errorf("BeegfsStorageCollector.Init(): BeeGFS file system statistics can only be queried by user root")
}
// Check if beegfs-ctl is in executable search path
_, err = exec.LookPath(m.config.Beegfs)
if err != nil {
return fmt.Errorf("BeegfsStorageCollector.Init(): Failed to find beegfs-ctl binary '%s': %v", m.config.Beegfs, err)
}
m.init = true
return nil
}
func (m *BeegfsStorageCollector) Read(interval time.Duration, output chan lp.CCMetric) {
if !m.init {
return
}
//get mounpoint
buffer, _ := ioutil.ReadFile(string("/proc/mounts"))
mounts := strings.Split(string(buffer), "\n")
var mountpoints []string
for _, line := range mounts {
if len(line) == 0 {
continue
}
f := strings.Fields(line)
if strings.Contains(f[0], "beegfs_ondemand") {
// Skip excluded filesystems
if _, skip := m.skipFS[f[1]]; skip {
continue
}
mountpoints = append(mountpoints, f[1])
}
}
if len(mountpoints) == 0 {
return
}
// collects stats for each BeeGFS on Demand FS
for _, mountpoint := range mountpoints {
m.tags["filesystem"] = mountpoint
// bwwgfs-ctl:
// --clientstats: Show client IO statistics.
// --nodetype=meta: The node type to query (meta, storage).
// --interval:
// --mount=/mnt/beeond/: Which mount point
//cmd := exec.Command(m.config.Beegfs, "/root/mc/test.txt")
mountoption := "--mount=" + mountpoint
cmd := exec.Command(m.config.Beegfs, "--clientstats",
"--nodetype=storage", mountoption, "--allstats")
cmd.Stdin = strings.NewReader("\n")
cmdStdout := new(bytes.Buffer)
cmdStderr := new(bytes.Buffer)
cmd.Stdout = cmdStdout
cmd.Stderr = cmdStderr
err := cmd.Run()
if err != nil {
fmt.Fprintf(os.Stderr, "BeegfsStorageCollector.Read(): Failed to execute command \"%s\": %s\n", cmd.String(), err.Error())
fmt.Fprintf(os.Stderr, "BeegfsStorageCollector.Read(): command exit code: \"%d\"\n", cmd.ProcessState.ExitCode())
data, _ := ioutil.ReadAll(cmdStderr)
fmt.Fprintf(os.Stderr, "BeegfsStorageCollector.Read(): command stderr: \"%s\"\n", string(data))
data, _ = ioutil.ReadAll(cmdStdout)
fmt.Fprintf(os.Stderr, "BeegfsStorageCollector.Read(): command stdout: \"%s\"\n", string(data))
return
}
// Read I/O statistics
scanner := bufio.NewScanner(cmdStdout)
sumLine := regexp.MustCompile(`^Sum:\s+\d+\s+\[[a-zA-Z]+\]+`)
//Line := regexp.MustCompile(`^(.*)\s+(\d)+\s+\[([a-zA-Z]+)\]+`)
statsLine := regexp.MustCompile(`^(.*?)\s+?(\d.*?)$`)
singleSpacePattern := regexp.MustCompile(`\s+`)
removePattern := regexp.MustCompile(`[\[|\]]`)
for scanner.Scan() {
readLine := scanner.Text()
//fmt.Println(readLine)
// Jump few lines, we only want the I/O stats from nodes
if !sumLine.MatchString(readLine) {
continue
}
match := statsLine.FindStringSubmatch(readLine)
// nodeName = "Sum:" or would be nodes
// nodeName := match[1]
//Remove multiple whitespaces
dummy := removePattern.ReplaceAllString(match[2], " ")
metaStats := strings.TrimSpace(singleSpacePattern.ReplaceAllString(dummy, " "))
split := strings.Split(metaStats, " ")
// fill map with values
// split[i+1] = mdname
// split[i] = amount of operations
for i := 0; i <= len(split)-1; i += 2 {
if _, ok := m.matches[split[i+1]]; ok {
m.matches["beegfs_cstorage_"+split[i+1]] = split[i]
//m.matches[split[i+1]] = split[i]
} else {
f1, err := strconv.ParseFloat(m.matches["other"], 32)
if err != nil {
cclog.ComponentError(
m.name,
fmt.Sprintf("Metric (other): Failed to convert str written '%s' to float: %v", m.matches["other"], err))
continue
}
f2, err := strconv.ParseFloat(split[i], 32)
if err != nil {
cclog.ComponentError(
m.name,
fmt.Sprintf("Metric (other): Failed to convert str written '%s' to float: %v", m.matches["other"], err))
continue
}
m.matches["beegfs_cstorage_other"] = fmt.Sprintf("%f", f1+f2)
}
}
for key, data := range m.matches {
value, _ := strconv.ParseFloat(data, 32)
y, err := lp.New(key, m.tags, m.meta, map[string]interface{}{"value": value}, time.Now())
if err == nil {
output <- y
}
}
}
}
}
func (m *BeegfsStorageCollector) Close() {
m.init = false
}

View File

@@ -0,0 +1,55 @@
## `BeeGFS on Demand` collector
This Collector is to collect BeeGFS on Demand (BeeOND) storage stats.
```json
"beegfs_storage": {
"beegfs_path": "/usr/bin/beegfs-ctl",
"exclude_filesystem": [
"/mnt/ignore_me"
],
"exclude_metrics": [
"ack",
"storInf",
"unlnk"
]
}
```
The `BeeGFS On Demand (BeeOND)` collector uses the `beegfs-ctl` command to read performance metrics for BeeGFS filesystems.
The reported filesystems can be filtered with the `exclude_filesystem` option
in the configuration.
The path to the `beegfs-ctl` command can be configured with the `beegfs_path` option
in the configuration.
When using the `exclude_metrics` option, the excluded metrics are summed as `other`.
Important: The metrics listed below, are similar to the naming of BeeGFS. The Collector prefixes these with `beegfs_cstorage_`(beegfs client meta).
For example beegfs metric `open`-> `beegfs_cstorage_`
Note: BeeGFS FS offers many Metadata Information. Probably it makes sense to exlcude most of them. Nevertheless, these excluded metrics will be summed as `beegfs_cstorage_other`.
Available Metrics:
* "sum"
* "ack"
* "sChDrct"
* "getFSize"
* "sAttr"
* "statfs"
* "trunc"
* "close"
* "fsync"
* "ops-rd"
* "MiB-rd/s"
* "ops-wr"
* "MiB-wr/s"
* "endbg"
* "hrtbeat"
* "remNode"
* "storInf"
* "unlnk"
The collector adds a `filesystem` tag to all metrics

View File

@@ -14,38 +14,43 @@ import (
// Map of all available metric collectors
var AvailableCollectors = map[string]MetricCollector{
"likwid": new(LikwidCollector),
"loadavg": new(LoadavgCollector),
"memstat": new(MemstatCollector),
"netstat": new(NetstatCollector),
"ibstat": new(InfinibandCollector),
"ibstat_perfquery": new(InfinibandPerfQueryCollector),
"lustrestat": new(LustreCollector),
"cpustat": new(CpustatCollector),
"topprocs": new(TopProcsCollector),
"nvidia": new(NvidiaCollector),
"customcmd": new(CustomCmdCollector),
"iostat": new(IOstatCollector),
"diskstat": new(DiskstatCollector),
"tempstat": new(TempCollector),
"ipmistat": new(IpmiCollector),
"gpfs": new(GpfsCollector),
"cpufreq": new(CPUFreqCollector),
"cpufreq_cpuinfo": new(CPUFreqCpuInfoCollector),
"nfs3stat": new(Nfs3Collector),
"nfs4stat": new(Nfs4Collector),
"numastats": new(NUMAStatsCollector),
"likwid": new(LikwidCollector),
"loadavg": new(LoadavgCollector),
"memstat": new(MemstatCollector),
"netstat": new(NetstatCollector),
"ibstat": new(InfinibandCollector),
"lustrestat": new(LustreCollector),
"cpustat": new(CpustatCollector),
"topprocs": new(TopProcsCollector),
"nvidia": new(NvidiaCollector),
"customcmd": new(CustomCmdCollector),
"iostat": new(IOstatCollector),
"diskstat": new(DiskstatCollector),
"tempstat": new(TempCollector),
"ipmistat": new(IpmiCollector),
"gpfs": new(GpfsCollector),
"cpufreq": new(CPUFreqCollector),
"cpufreq_cpuinfo": new(CPUFreqCpuInfoCollector),
"nfs3stat": new(Nfs3Collector),
"nfs4stat": new(Nfs4Collector),
"numastats": new(NUMAStatsCollector),
"beegfs_meta": new(BeegfsMetaCollector),
"beegfs_storage": new(BeegfsStorageCollector),
"rocm_smi": new(RocmSmiCollector),
}
// Metric collector manager data structure
type collectorManager struct {
collectors []MetricCollector // List of metric collectors to use
output chan lp.CCMetric // Output channels
done chan bool // channel to finish / stop metric collector manager
ticker mct.MultiChanTicker // periodically ticking once each interval
duration time.Duration // duration (for metrics that measure over a given duration)
wg *sync.WaitGroup // wait group for all goroutines in cc-metric-collector
config map[string]json.RawMessage // json encoded config for collector manager
collectors []MetricCollector // List of metric collectors to read in parallel
serial []MetricCollector // List of metric collectors to read serially
output chan lp.CCMetric // Output channels
done chan bool // channel to finish / stop metric collector manager
ticker mct.MultiChanTicker // periodically ticking once each interval
duration time.Duration // duration (for metrics that measure over a given duration)
wg *sync.WaitGroup // wait group for all goroutines in cc-metric-collector
config map[string]json.RawMessage // json encoded config for collector manager
collector_wg sync.WaitGroup // internally used wait group for the parallel reading of collector
parallel_run bool // Flag whether the collectors are currently read in parallel
}
// Metric collector manager access functions
@@ -65,6 +70,7 @@ type CollectorManager interface {
// Initialization is done for all configured collectors
func (cm *collectorManager) Init(ticker mct.MultiChanTicker, duration time.Duration, wg *sync.WaitGroup, collectConfigFile string) error {
cm.collectors = make([]MetricCollector, 0)
cm.serial = make([]MetricCollector, 0)
cm.output = nil
cm.done = make(chan bool)
cm.wg = wg
@@ -99,7 +105,11 @@ func (cm *collectorManager) Init(ticker mct.MultiChanTicker, duration time.Durat
continue
}
cclog.ComponentDebug("CollectorManager", "ADD COLLECTOR", collector.Name())
cm.collectors = append(cm.collectors, collector)
if collector.Parallel() {
cm.collectors = append(cm.collectors, collector)
} else {
cm.serial = append(cm.serial, collector)
}
}
return nil
}
@@ -115,6 +125,10 @@ func (cm *collectorManager) Start() {
// Collector manager is done
done := func() {
// close all metric collectors
if cm.parallel_run {
cm.collector_wg.Wait()
cm.parallel_run = false
}
for _, c := range cm.collectors {
c.Close()
}
@@ -129,7 +143,26 @@ func (cm *collectorManager) Start() {
done()
return
case t := <-tick:
cm.parallel_run = true
for _, c := range cm.collectors {
// Wait for done signal or execute the collector
select {
case <-cm.done:
done()
return
default:
// Read metrics from collector c via goroutine
cclog.ComponentDebug("CollectorManager", c.Name(), t)
cm.collector_wg.Add(1)
go func(myc MetricCollector) {
myc.Read(cm.duration, cm.output)
cm.collector_wg.Done()
}(c)
}
}
cm.collector_wg.Wait()
cm.parallel_run = false
for _, c := range cm.serial {
// Wait for done signal or execute the collector
select {
case <-cm.done:

View File

@@ -48,6 +48,7 @@ func (m *CPUFreqCpuInfoCollector) Init(config json.RawMessage) error {
m.setup()
m.name = "CPUFreqCpuInfoCollector"
m.parallel = true
m.meta = map[string]string{
"source": m.name,
"group": "CPU",
@@ -57,7 +58,7 @@ func (m *CPUFreqCpuInfoCollector) Init(config json.RawMessage) error {
const cpuInfoFile = "/proc/cpuinfo"
file, err := os.Open(cpuInfoFile)
if err != nil {
return fmt.Errorf("Failed to open file '%s': %v", cpuInfoFile, err)
return fmt.Errorf("failed to open file '%s': %v", cpuInfoFile, err)
}
defer file.Close()
@@ -106,14 +107,14 @@ func (m *CPUFreqCpuInfoCollector) Init(config json.RawMessage) error {
topology.coreID = coreID
topology.coreID_int, err = strconv.ParseInt(coreID, 10, 64)
if err != nil {
return fmt.Errorf("Unable to convert coreID '%s' to int64: %v", coreID, err)
return fmt.Errorf("unable to convert coreID '%s' to int64: %v", coreID, err)
}
// Physical package ID
topology.physicalPackageID = physicalPackageID
topology.physicalPackageID_int, err = strconv.ParseInt(physicalPackageID, 10, 64)
if err != nil {
return fmt.Errorf("Unable to convert physicalPackageID '%s' to int64: %v", physicalPackageID, err)
return fmt.Errorf("unable to convert physicalPackageID '%s' to int64: %v", physicalPackageID, err)
}
// increase maximun socket / package ID, when required
@@ -150,7 +151,7 @@ func (m *CPUFreqCpuInfoCollector) Init(config json.RawMessage) error {
t.numNonHT = numNonHT
t.numNonHT_int = numNonHT_int
t.tagSet = map[string]string{
"type": "cpu",
"type": "hwthread",
"type-id": t.processor,
"package_id": t.physicalPackageID,
}

View File

@@ -4,7 +4,7 @@
"cpufreq_cpuinfo": {}
```
The `cpufreq_cpuinfo` collector reads the clock frequency from `/proc/cpuinfo` and outputs a handful **cpu** metrics.
The `cpufreq_cpuinfo` collector reads the clock frequency from `/proc/cpuinfo` and outputs a handful **hwthread** metrics.
Metrics:
* `cpufreq`

View File

@@ -53,6 +53,7 @@ func (m *CPUFreqCollector) Init(config json.RawMessage) error {
m.name = "CPUFreqCollector"
m.setup()
m.parallel = true
if len(config) > 0 {
err := json.Unmarshal(config, &m.config)
if err != nil {
@@ -70,10 +71,10 @@ func (m *CPUFreqCollector) Init(config json.RawMessage) error {
globPattern := filepath.Join(baseDir, "cpu[0-9]*")
cpuDirs, err := filepath.Glob(globPattern)
if err != nil {
return fmt.Errorf("Unable to glob files with pattern '%s': %v", globPattern, err)
return fmt.Errorf("unable to glob files with pattern '%s': %v", globPattern, err)
}
if cpuDirs == nil {
return fmt.Errorf("Unable to find any files with pattern '%s'", globPattern)
return fmt.Errorf("unable to find any files with pattern '%s'", globPattern)
}
// Initialize CPU topology
@@ -82,38 +83,38 @@ func (m *CPUFreqCollector) Init(config json.RawMessage) error {
processor := strings.TrimPrefix(cpuDir, "/sys/devices/system/cpu/cpu")
processor_int, err := strconv.ParseInt(processor, 10, 64)
if err != nil {
return fmt.Errorf("Unable to convert cpuID '%s' to int64: %v", processor, err)
return fmt.Errorf("unable to convert cpuID '%s' to int64: %v", processor, err)
}
// Read package ID
physicalPackageIDFile := filepath.Join(cpuDir, "topology", "physical_package_id")
line, err := ioutil.ReadFile(physicalPackageIDFile)
if err != nil {
return fmt.Errorf("Unable to read physical package ID from file '%s': %v", physicalPackageIDFile, err)
return fmt.Errorf("unable to read physical package ID from file '%s': %v", physicalPackageIDFile, err)
}
physicalPackageID := strings.TrimSpace(string(line))
physicalPackageID_int, err := strconv.ParseInt(physicalPackageID, 10, 64)
if err != nil {
return fmt.Errorf("Unable to convert packageID '%s' to int64: %v", physicalPackageID, err)
return fmt.Errorf("unable to convert packageID '%s' to int64: %v", physicalPackageID, err)
}
// Read core ID
coreIDFile := filepath.Join(cpuDir, "topology", "core_id")
line, err = ioutil.ReadFile(coreIDFile)
if err != nil {
return fmt.Errorf("Unable to read core ID from file '%s': %v", coreIDFile, err)
return fmt.Errorf("unable to read core ID from file '%s': %v", coreIDFile, err)
}
coreID := strings.TrimSpace(string(line))
coreID_int, err := strconv.ParseInt(coreID, 10, 64)
if err != nil {
return fmt.Errorf("Unable to convert coreID '%s' to int64: %v", coreID, err)
return fmt.Errorf("unable to convert coreID '%s' to int64: %v", coreID, err)
}
// Check access to current frequency file
scalingCurFreqFile := filepath.Join(cpuDir, "cpufreq", "scaling_cur_freq")
err = unix.Access(scalingCurFreqFile, unix.R_OK)
if err != nil {
return fmt.Errorf("Unable to access file '%s': %v", scalingCurFreqFile, err)
return fmt.Errorf("unable to access file '%s': %v", scalingCurFreqFile, err)
}
t := &m.topology[processor_int]
@@ -161,7 +162,7 @@ func (m *CPUFreqCollector) Init(config json.RawMessage) error {
t.numNonHT = numNonHT
t.numNonHT_int = numNonHT_int
t.tagSet = map[string]string{
"type": "cpu",
"type": "hwthread",
"type-id": t.processor,
"package_id": t.physicalPackageID,
}

View File

@@ -5,7 +5,7 @@
}
```
The `cpufreq` collector reads the clock frequency from `/sys/devices/system/cpu/cpu*/cpufreq` and outputs a handful **cpu** metrics.
The `cpufreq` collector reads the clock frequency from `/sys/devices/system/cpu/cpu*/cpufreq` and outputs a handful **hwthread** metrics.
Metrics:
* `cpufreq`

View File

@@ -21,16 +21,16 @@ type CpustatCollectorConfig struct {
type CpustatCollector struct {
metricCollector
config CpustatCollectorConfig
matches map[string]int
cputags map[string]map[string]string
nodetags map[string]string
num_cpus_metric lp.CCMetric
config CpustatCollectorConfig
matches map[string]int
cputags map[string]map[string]string
nodetags map[string]string
}
func (m *CpustatCollector) Init(config json.RawMessage) error {
m.name = "CpustatCollector"
m.setup()
m.parallel = true
m.meta = map[string]string{"source": m.name, "group": "CPU", "unit": "Percent"}
m.nodetags = map[string]string{"type": "node"}
if len(config) > 0 {
@@ -83,7 +83,7 @@ func (m *CpustatCollector) Init(config json.RawMessage) error {
if strings.HasPrefix(linefields[0], "cpu") && strings.Compare(linefields[0], "cpu") != 0 {
cpustr := strings.TrimLeft(linefields[0], "cpu")
cpu, _ := strconv.Atoi(cpustr)
m.cputags[linefields[0]] = map[string]string{"type": "cpu", "type-id": fmt.Sprintf("%d", cpu)}
m.cputags[linefields[0]] = map[string]string{"type": "hwthread", "type-id": fmt.Sprintf("%d", cpu)}
num_cpus++
}
}

View File

@@ -33,6 +33,7 @@ type CustomCmdCollector struct {
func (m *CustomCmdCollector) Init(config json.RawMessage) error {
var err error
m.name = "CustomCmdCollector"
m.parallel = true
m.meta = map[string]string{"source": m.name, "group": "Custom"}
if len(config) > 0 {
err = json.Unmarshal(config, &m.config)
@@ -61,7 +62,7 @@ func (m *CustomCmdCollector) Init(config json.RawMessage) error {
}
}
if len(m.files) == 0 && len(m.commands) == 0 {
return errors.New("No metrics to collect")
return errors.New("no metrics to collect")
}
m.handler = influx.NewMetricHandler()
m.parser = influx.NewParser(m.handler)

View File

@@ -3,7 +3,6 @@ package collectors
import (
"bufio"
"encoding/json"
"fmt"
"os"
"strings"
"syscall"
@@ -30,6 +29,7 @@ type DiskstatCollector struct {
func (m *DiskstatCollector) Init(config json.RawMessage) error {
m.name = "DiskstatCollector"
m.parallel = true
m.meta = map[string]string{"source": m.name, "group": "Disk"}
m.setup()
if len(config) > 0 {
@@ -78,11 +78,17 @@ func (m *DiskstatCollector) Read(interval time.Duration, output chan lp.CCMetric
continue
}
path := strings.Replace(linefields[1], `\040`, " ", -1)
stat := syscall.Statfs_t{}
stat := syscall.Statfs_t{
Blocks: 0,
Bsize: 0,
Bfree: 0,
}
err := syscall.Statfs(path, &stat)
if err != nil {
fmt.Println(err.Error())
return
continue
}
if stat.Blocks == 0 || stat.Bsize == 0 {
continue
}
tags := map[string]string{"type": "node", "device": linefields[0]}
total := (stat.Blocks * uint64(stat.Bsize)) / uint64(1000000000)
@@ -97,9 +103,11 @@ func (m *DiskstatCollector) Read(interval time.Duration, output chan lp.CCMetric
y.AddMeta("unit", "GBytes")
output <- y
}
perc := (100 * (total - free)) / total
if perc > part_max_used {
part_max_used = perc
if total > 0 {
perc := (100 * (total - free)) / total
if perc > part_max_used {
part_max_used = perc
}
}
}
y, err := lp.New("part_max_used", map[string]string{"type": "node"}, m.meta, map[string]interface{}{"value": int(part_max_used)}, time.Now())

View File

@@ -17,14 +17,24 @@ import (
lp "github.com/ClusterCockpit/cc-metric-collector/internal/ccMetric"
)
const DEFAULT_GPFS_CMD = "mmpmon"
type GpfsCollectorLastState struct {
bytesRead int64
bytesWritten int64
}
type GpfsCollector struct {
metricCollector
tags map[string]string
config struct {
Mmpmon string `json:"mmpmon_path,omitempty"`
ExcludeFilesystem []string `json:"exclude_filesystem,omitempty"`
SendBandwidths bool `json:"send_bandwidths"`
}
skipFS map[string]struct{}
skipFS map[string]struct{}
lastTimestamp time.Time // Store time stamp of last tick to derive bandwidths
lastState map[string]GpfsCollectorLastState
}
func (m *GpfsCollector) Init(config json.RawMessage) error {
@@ -36,9 +46,10 @@ func (m *GpfsCollector) Init(config json.RawMessage) error {
var err error
m.name = "GpfsCollector"
m.setup()
m.parallel = true
// Set default mmpmon binary
m.config.Mmpmon = "/usr/lpp/mmfs/bin/mmpmon"
m.config.Mmpmon = DEFAULT_GPFS_CMD
// Read JSON configuration
if len(config) > 0 {
@@ -60,21 +71,23 @@ func (m *GpfsCollector) Init(config json.RawMessage) error {
for _, fs := range m.config.ExcludeFilesystem {
m.skipFS[fs] = struct{}{}
}
m.lastState = make(map[string]GpfsCollectorLastState)
// GPFS / IBM Spectrum Scale file system statistics can only be queried by user root
user, err := user.Current()
if err != nil {
return fmt.Errorf("Failed to get current user: %v", err)
return fmt.Errorf("failed to get current user: %v", err)
}
if user.Uid != "0" {
return fmt.Errorf("GPFS file system statistics can only be queried by user root")
}
// Check if mmpmon is in executable search path
_, err = exec.LookPath(m.config.Mmpmon)
p, err := exec.LookPath(m.config.Mmpmon)
if err != nil {
return fmt.Errorf("Failed to find mmpmon binary '%s': %v", m.config.Mmpmon, err)
return fmt.Errorf("failed to find mmpmon binary '%s': %v", m.config.Mmpmon, err)
}
m.config.Mmpmon = p
m.init = true
return nil
@@ -86,6 +99,13 @@ func (m *GpfsCollector) Read(interval time.Duration, output chan lp.CCMetric) {
return
}
// Current time stamp
now := time.Now()
// time difference to last time stamp
timeDiff := now.Sub(m.lastTimestamp).Seconds()
// Save current timestamp
m.lastTimestamp = now
// mmpmon:
// -p: generate output that can be parsed
// -s: suppress the prompt on input
@@ -144,8 +164,19 @@ func (m *GpfsCollector) Read(interval time.Duration, output chan lp.CCMetric) {
continue
}
// Add filesystem tag
m.tags["filesystem"] = filesystem
// Create initial last state
if m.config.SendBandwidths {
if _, ok := m.lastState[filesystem]; !ok {
m.lastState[filesystem] = GpfsCollectorLastState{
bytesRead: -1,
bytesWritten: -1,
}
}
}
// return code
rc, err := strconv.Atoi(key_value["_rc_"])
if err != nil {
@@ -188,6 +219,14 @@ func (m *GpfsCollector) Read(interval time.Duration, output chan lp.CCMetric) {
if y, err := lp.New("gpfs_bytes_read", m.tags, m.meta, map[string]interface{}{"value": bytesRead}, timestamp); err == nil {
output <- y
}
if m.config.SendBandwidths {
if lastBytesRead := m.lastState[filesystem].bytesRead; lastBytesRead >= 0 {
bwRead := float64(bytesRead-lastBytesRead) / timeDiff
if y, err := lp.New("gpfs_bw_read", m.tags, m.meta, map[string]interface{}{"value": bwRead}, timestamp); err == nil {
output <- y
}
}
}
// bytes written
bytesWritten, err := strconv.ParseInt(key_value["_bw_"], 10, 64)
@@ -200,6 +239,21 @@ func (m *GpfsCollector) Read(interval time.Duration, output chan lp.CCMetric) {
if y, err := lp.New("gpfs_bytes_written", m.tags, m.meta, map[string]interface{}{"value": bytesWritten}, timestamp); err == nil {
output <- y
}
if m.config.SendBandwidths {
if lastBytesWritten := m.lastState[filesystem].bytesRead; lastBytesWritten >= 0 {
bwWrite := float64(bytesWritten-lastBytesWritten) / timeDiff
if y, err := lp.New("gpfs_bw_write", m.tags, m.meta, map[string]interface{}{"value": bwWrite}, timestamp); err == nil {
output <- y
}
}
}
if m.config.SendBandwidths {
m.lastState[filesystem] = GpfsCollectorLastState{
bytesRead: bytesRead,
bytesWritten: bytesWritten,
}
}
// number of opens
numOpens, err := strconv.ParseInt(key_value["_oc_"], 10, 64)

View File

@@ -5,7 +5,8 @@
"mmpmon_path": "/path/to/mmpmon",
"exclude_filesystem": [
"fs1"
]
],
"send_bandwidths" : true
}
```
@@ -16,15 +17,18 @@ The reported filesystems can be filtered with the `exclude_filesystem` option
in the configuration.
The path to the `mmpmon` command can be configured with the `mmpmon_path` option
in the configuration.
in the configuration. If nothing is set, the collector searches in `$PATH` for `mmpmon`.
Metrics:
* `bytes_read`
* `gpfs_bytes_read`
* `gpfs_bytes_written`
* `gpfs_num_opens`
* `gpfs_num_closes`
* `gpfs_num_reads`
* `gpfs_num_readdirs`
* `gpfs_num_inode_updates`
* `gpfs_bw_read` (if `send_bandwidths == true`)
* `gpfs_bw_write` (if `send_bandwidths == true`)
The collector adds a `filesystem` tag to all metrics

View File

@@ -16,22 +16,31 @@ import (
"time"
)
const IB_BASEPATH = `/sys/class/infiniband/`
const IB_BASEPATH = "/sys/class/infiniband/"
type InfinibandCollectorMetric struct {
path string
unit string
}
type InfinibandCollectorInfo struct {
LID string // IB local Identifier (LID)
device string // IB device
port string // IB device port
portCounterFiles map[string]string // mapping counter name -> sysfs file
tagSet map[string]string // corresponding tag list
LID string // IB local Identifier (LID)
device string // IB device
port string // IB device port
portCounterFiles map[string]InfinibandCollectorMetric // mapping counter name -> InfinibandCollectorMetric
tagSet map[string]string // corresponding tag list
lastState map[string]int64 // State from last measurement
}
type InfinibandCollector struct {
metricCollector
config struct {
ExcludeDevices []string `json:"exclude_devices,omitempty"` // IB device to exclude e.g. mlx5_0
ExcludeDevices []string `json:"exclude_devices,omitempty"` // IB device to exclude e.g. mlx5_0
SendAbsoluteValues bool `json:"send_abs_values"` // Send absolut values as read from sys filesystem
SendDerivedValues bool `json:"send_derived_values"` // Send derived values e.g. rates
}
info []*InfinibandCollectorInfo
info []*InfinibandCollectorInfo
lastTimestamp time.Time // Store time stamp of last tick to derive bandwidths
}
// Init initializes the Infiniband collector by walking through files below IB_BASEPATH
@@ -45,10 +54,16 @@ func (m *InfinibandCollector) Init(config json.RawMessage) error {
var err error
m.name = "InfinibandCollector"
m.setup()
m.parallel = true
m.meta = map[string]string{
"source": m.name,
"group": "Network",
}
// Set default configuration,
m.config.SendAbsoluteValues = true
m.config.SendDerivedValues = false
// Read configuration file, allow overwriting default config
if len(config) > 0 {
err = json.Unmarshal(config, &m.config)
if err != nil {
@@ -60,10 +75,10 @@ func (m *InfinibandCollector) Init(config json.RawMessage) error {
globPattern := filepath.Join(IB_BASEPATH, "*", "ports", "*")
ibDirs, err := filepath.Glob(globPattern)
if err != nil {
return fmt.Errorf("Unable to glob files with pattern %s: %v", globPattern, err)
return fmt.Errorf("unable to glob files with pattern %s: %v", globPattern, err)
}
if ibDirs == nil {
return fmt.Errorf("Unable to find any directories with pattern %s", globPattern)
return fmt.Errorf("unable to find any directories with pattern %s", globPattern)
}
for _, path := range ibDirs {
@@ -97,19 +112,25 @@ func (m *InfinibandCollector) Init(config json.RawMessage) error {
// Check access to counter files
countersDir := filepath.Join(path, "counters")
portCounterFiles := map[string]string{
"ib_recv": filepath.Join(countersDir, "port_rcv_data"),
"ib_xmit": filepath.Join(countersDir, "port_xmit_data"),
"ib_recv_pkts": filepath.Join(countersDir, "port_rcv_packets"),
"ib_xmit_pkts": filepath.Join(countersDir, "port_xmit_packets"),
portCounterFiles := map[string]InfinibandCollectorMetric{
"ib_recv": {path: filepath.Join(countersDir, "port_rcv_data"), unit: "bytes"},
"ib_xmit": {path: filepath.Join(countersDir, "port_xmit_data"), unit: "bytes"},
"ib_recv_pkts": {path: filepath.Join(countersDir, "port_rcv_packets"), unit: "packets"},
"ib_xmit_pkts": {path: filepath.Join(countersDir, "port_xmit_packets"), unit: "packets"},
}
for _, counterFile := range portCounterFiles {
err := unix.Access(counterFile, unix.R_OK)
for _, counter := range portCounterFiles {
err := unix.Access(counter.path, unix.R_OK)
if err != nil {
return fmt.Errorf("Unable to access %s: %v", counterFile, err)
return fmt.Errorf("unable to access %s: %v", counter.path, err)
}
}
// Initialize last state
lastState := make(map[string]int64)
for counter := range portCounterFiles {
lastState[counter] = -1
}
m.info = append(m.info,
&InfinibandCollectorInfo{
LID: LID,
@@ -122,11 +143,12 @@ func (m *InfinibandCollector) Init(config json.RawMessage) error {
"port": port,
"lid": LID,
},
lastState: lastState,
})
}
if len(m.info) == 0 {
return fmt.Errorf("Found no IB devices")
return fmt.Errorf("found no IB devices")
}
m.init = true
@@ -141,17 +163,27 @@ func (m *InfinibandCollector) Read(interval time.Duration, output chan lp.CCMetr
return
}
// Current time stamp
now := time.Now()
// time difference to last time stamp
timeDiff := now.Sub(m.lastTimestamp).Seconds()
// Save current timestamp
m.lastTimestamp = now
for _, info := range m.info {
for counterName, counterFile := range info.portCounterFiles {
line, err := ioutil.ReadFile(counterFile)
for counterName, counterDef := range info.portCounterFiles {
// Read counter file
line, err := ioutil.ReadFile(counterDef.path)
if err != nil {
cclog.ComponentError(
m.name,
fmt.Sprintf("Read(): Failed to read from file '%s': %v", counterFile, err))
fmt.Sprintf("Read(): Failed to read from file '%s': %v", counterDef.path, err))
continue
}
data := strings.TrimSpace(string(line))
// convert counter to int64
v, err := strconv.ParseInt(data, 10, 64)
if err != nil {
cclog.ComponentError(
@@ -159,8 +191,26 @@ func (m *InfinibandCollector) Read(interval time.Duration, output chan lp.CCMetr
fmt.Sprintf("Read(): Failed to convert Infininiband metrice %s='%s' to int64: %v", counterName, data, err))
continue
}
if y, err := lp.New(counterName, info.tagSet, m.meta, map[string]interface{}{"value": v}, now); err == nil {
output <- y
// Send absolut values
if m.config.SendAbsoluteValues {
if y, err := lp.New(counterName, info.tagSet, m.meta, map[string]interface{}{"value": v}, now); err == nil {
y.AddMeta("unit", counterDef.unit)
output <- y
}
}
// Send derived values
if m.config.SendDerivedValues {
if info.lastState[counterName] >= 0 {
rate := float64((v - info.lastState[counterName])) / timeDiff
if y, err := lp.New(counterName+"_bw", info.tagSet, m.meta, map[string]interface{}{"value": rate}, now); err == nil {
y.AddMeta("unit", counterDef.unit+"/sec")
output <- y
}
}
// Save current state
info.lastState[counterName] = v
}
}

View File

@@ -5,7 +5,9 @@
"ibstat": {
"exclude_devices": [
"mlx4"
]
],
"send_abs_values": true,
"send_derived_values": true
}
```
@@ -22,5 +24,9 @@ Metrics:
* `ib_xmit`
* `ib_recv_pkts`
* `ib_xmit_pkts`
* `ib_recv_bw` (if `send_derived_values == true`)
* `ib_xmit_bw` (if `send_derived_values == true`)
* `ib_recv_pkts_bw` (if `send_derived_values == true`)
* `ib_xmit_pkts_bw` (if `send_derived_values == true`)
The collector adds a `device` tag to all metrics

View File

@@ -1,232 +0,0 @@
package collectors
import (
"fmt"
"io/ioutil"
"log"
"os/exec"
lp "github.com/ClusterCockpit/cc-metric-collector/internal/ccMetric"
// "os"
"encoding/json"
"errors"
"path/filepath"
"strconv"
"strings"
"time"
)
const PERFQUERY = `/usr/sbin/perfquery`
type InfinibandPerfQueryCollector struct {
metricCollector
tags map[string]string
lids map[string]map[string]string
config struct {
ExcludeDevices []string `json:"exclude_devices,omitempty"`
PerfQueryPath string `json:"perfquery_path"`
}
}
func (m *InfinibandPerfQueryCollector) Init(config json.RawMessage) error {
var err error
m.name = "InfinibandCollectorPerfQuery"
m.setup()
m.meta = map[string]string{"source": m.name, "group": "Network"}
m.tags = map[string]string{"type": "node"}
if len(config) > 0 {
err = json.Unmarshal(config, &m.config)
if err != nil {
return err
}
}
if len(m.config.PerfQueryPath) == 0 {
path, err := exec.LookPath("perfquery")
if err == nil {
m.config.PerfQueryPath = path
}
}
m.lids = make(map[string]map[string]string)
p := fmt.Sprintf("%s/*/ports/*/lid", string(IB_BASEPATH))
files, err := filepath.Glob(p)
if err != nil {
return err
}
for _, f := range files {
lid, err := ioutil.ReadFile(f)
if err == nil {
plist := strings.Split(strings.Replace(f, string(IB_BASEPATH), "", -1), "/")
skip := false
for _, d := range m.config.ExcludeDevices {
if d == plist[0] {
skip = true
}
}
if !skip {
m.lids[plist[0]] = make(map[string]string)
m.lids[plist[0]][plist[2]] = string(lid)
}
}
}
for _, ports := range m.lids {
for port, lid := range ports {
args := fmt.Sprintf("-r %s %s 0xf000", lid, port)
command := exec.Command(m.config.PerfQueryPath, args)
command.Wait()
_, err := command.Output()
if err != nil {
return fmt.Errorf("Failed to execute %s: %v", m.config.PerfQueryPath, err)
}
}
}
if len(m.lids) == 0 {
return errors.New("No usable IB devices")
}
m.init = true
return nil
}
func (m *InfinibandPerfQueryCollector) doPerfQuery(cmd string, dev string, lid string, port string, tags map[string]string, output chan lp.CCMetric) error {
args := fmt.Sprintf("-r %s %s 0xf000", lid, port)
command := exec.Command(cmd, args)
command.Wait()
stdout, err := command.Output()
if err != nil {
log.Print(err)
return err
}
ll := strings.Split(string(stdout), "\n")
for _, line := range ll {
if strings.HasPrefix(line, "PortRcvData") || strings.HasPrefix(line, "RcvData") {
lv := strings.Fields(line)
v, err := strconv.ParseFloat(lv[1], 64)
if err == nil {
y, err := lp.New("ib_recv", tags, m.meta, map[string]interface{}{"value": float64(v)}, time.Now())
if err == nil {
output <- y
}
}
}
if strings.HasPrefix(line, "PortXmitData") || strings.HasPrefix(line, "XmtData") {
lv := strings.Fields(line)
v, err := strconv.ParseFloat(lv[1], 64)
if err == nil {
y, err := lp.New("ib_xmit", tags, m.meta, map[string]interface{}{"value": float64(v)}, time.Now())
if err == nil {
output <- y
}
}
}
if strings.HasPrefix(line, "PortRcvPkts") || strings.HasPrefix(line, "RcvPkts") {
lv := strings.Fields(line)
v, err := strconv.ParseFloat(lv[1], 64)
if err == nil {
y, err := lp.New("ib_recv_pkts", tags, m.meta, map[string]interface{}{"value": float64(v)}, time.Now())
if err == nil {
output <- y
}
}
}
if strings.HasPrefix(line, "PortXmitPkts") || strings.HasPrefix(line, "XmtPkts") {
lv := strings.Fields(line)
v, err := strconv.ParseFloat(lv[1], 64)
if err == nil {
y, err := lp.New("ib_xmit_pkts", tags, m.meta, map[string]interface{}{"value": float64(v)}, time.Now())
if err == nil {
output <- y
}
}
}
if strings.HasPrefix(line, "PortRcvPkts") || strings.HasPrefix(line, "RcvPkts") {
lv := strings.Fields(line)
v, err := strconv.ParseFloat(lv[1], 64)
if err == nil {
y, err := lp.New("ib_recv_pkts", tags, m.meta, map[string]interface{}{"value": float64(v)}, time.Now())
if err == nil {
output <- y
}
}
}
if strings.HasPrefix(line, "PortXmitPkts") || strings.HasPrefix(line, "XmtPkts") {
lv := strings.Fields(line)
v, err := strconv.ParseFloat(lv[1], 64)
if err == nil {
y, err := lp.New("ib_xmit_pkts", tags, m.meta, map[string]interface{}{"value": float64(v)}, time.Now())
if err == nil {
output <- y
}
}
}
}
return nil
}
func (m *InfinibandPerfQueryCollector) Read(interval time.Duration, output chan lp.CCMetric) {
if m.init {
for dev, ports := range m.lids {
for port, lid := range ports {
tags := map[string]string{
"type": "node",
"device": dev,
"port": port,
"lid": lid}
path := fmt.Sprintf("%s/%s/ports/%s/counters/", string(IB_BASEPATH), dev, port)
buffer, err := ioutil.ReadFile(fmt.Sprintf("%s/port_rcv_data", path))
if err == nil {
data := strings.Replace(string(buffer), "\n", "", -1)
v, err := strconv.ParseFloat(data, 64)
if err == nil {
y, err := lp.New("ib_recv", tags, m.meta, map[string]interface{}{"value": float64(v)}, time.Now())
if err == nil {
output <- y
}
}
}
buffer, err = ioutil.ReadFile(fmt.Sprintf("%s/port_xmit_data", path))
if err == nil {
data := strings.Replace(string(buffer), "\n", "", -1)
v, err := strconv.ParseFloat(data, 64)
if err == nil {
y, err := lp.New("ib_xmit", tags, m.meta, map[string]interface{}{"value": float64(v)}, time.Now())
if err == nil {
output <- y
}
}
}
buffer, err = ioutil.ReadFile(fmt.Sprintf("%s/port_rcv_packets", path))
if err == nil {
data := strings.Replace(string(buffer), "\n", "", -1)
v, err := strconv.ParseFloat(data, 64)
if err == nil {
y, err := lp.New("ib_recv_pkts", tags, m.meta, map[string]interface{}{"value": float64(v)}, time.Now())
if err == nil {
output <- y
}
}
}
buffer, err = ioutil.ReadFile(fmt.Sprintf("%s/port_xmit_packets", path))
if err == nil {
data := strings.Replace(string(buffer), "\n", "", -1)
v, err := strconv.ParseFloat(data, 64)
if err == nil {
y, err := lp.New("ib_xmit_pkts", tags, m.meta, map[string]interface{}{"value": float64(v)}, time.Now())
if err == nil {
output <- y
}
}
}
}
}
}
}
func (m *InfinibandPerfQueryCollector) Close() {
m.init = false
}

View File

@@ -1,28 +0,0 @@
## `ibstat_perfquery` collector
```json
"ibstat_perfquery": {
"perfquery_path": "/path/to/perfquery",
"exclude_devices": [
"mlx4"
]
}
```
The `ibstat_perfquery` collector includes all Infiniband devices that can be
found below `/sys/class/infiniband/` and where any of the ports provides a
LID file (`/sys/class/infiniband/<dev>/ports/<port>/lid`)
The devices can be filtered with the `exclude_devices` option in the configuration.
For each found LID the collector calls the `perfquery` command. The path to the
`perfquery` command can be configured with the `perfquery_path` option in the configuration
Metrics:
* `ib_recv`
* `ib_xmit`
* `ib_recv_pkts`
* `ib_xmit_pkts`
The collector adds a `device` tag to all metrics

View File

@@ -37,6 +37,7 @@ type IOstatCollector struct {
func (m *IOstatCollector) Init(config json.RawMessage) error {
var err error
m.name = "IOstatCollector"
m.parallel = true
m.meta = map[string]string{"source": m.name, "group": "Disk"}
m.setup()
if len(config) > 0 {

View File

@@ -34,6 +34,7 @@ type IpmiCollector struct {
func (m *IpmiCollector) Init(config json.RawMessage) error {
m.name = "IpmiCollector"
m.setup()
m.parallel = true
m.meta = map[string]string{"source": m.name, "group": "IPMI"}
m.config.IpmitoolPath = string(IPMITOOL_PATH)
m.config.IpmisensorsPath = string(IPMISENSORS_PATH)
@@ -54,7 +55,7 @@ func (m *IpmiCollector) Init(config json.RawMessage) error {
m.ipmisensors = p
}
if len(m.ipmitool) == 0 && len(m.ipmisensors) == 0 {
return errors.New("No IPMI reader found")
return errors.New("no IPMI reader found")
}
m.init = true
return nil

View File

@@ -15,9 +15,12 @@ import (
"io/ioutil"
"math"
"os"
"regexp"
"os/signal"
"sort"
"strconv"
"strings"
"sync"
"syscall"
"time"
"unsafe"
@@ -28,48 +31,6 @@ import (
"github.com/NVIDIA/go-nvml/pkg/dl"
)
type MetricScope string
const (
METRIC_SCOPE_HWTHREAD = iota
METRIC_SCOPE_CORE
METRIC_SCOPE_LLC
METRIC_SCOPE_NUMA
METRIC_SCOPE_DIE
METRIC_SCOPE_SOCKET
METRIC_SCOPE_NODE
)
func (ms MetricScope) String() string {
return string(ms)
}
func (ms MetricScope) Likwid() string {
LikwidDomains := map[string]string{
"cpu": "",
"core": "",
"llc": "C",
"numadomain": "M",
"die": "D",
"socket": "S",
"node": "N",
}
return LikwidDomains[string(ms)]
}
func (ms MetricScope) Granularity() int {
for i, g := range GetAllMetricScopes() {
if ms == g {
return i
}
}
return -1
}
func GetAllMetricScopes() []MetricScope {
return []MetricScope{"cpu" /*, "core", "llc", "numadomain", "die",*/, "socket", "node"}
}
const (
LIKWID_LIB_NAME = "liblikwid.so"
LIKWID_LIB_DL_FLAGS = dl.RTLD_LAZY | dl.RTLD_GLOBAL
@@ -77,18 +38,26 @@ const (
)
type LikwidCollectorMetricConfig struct {
Name string `json:"name"` // Name of the metric
Calc string `json:"calc"` // Calculation for the metric using
//Aggr string `json:"aggregation"` // if scope unequal to LIKWID metric scope, the values are combined (sum, min, max, mean or avg, median)
Scope MetricScope `json:"scope"` // scope for calculation. subscopes are aggregated using the 'aggregation' function
Publish bool `json:"publish"`
granulatity MetricScope
Name string `json:"name"` // Name of the metric
Calc string `json:"calc"` // Calculation for the metric using
Type string `json:"type"` // Metric type (aka node, socket, cpu, ...)
Publish bool `json:"publish"`
Unit string `json:"unit"` // Unit of metric if any
}
type LikwidCollectorEventsetConfig struct {
Events map[string]string `json:"events"`
granulatity map[string]MetricScope
Metrics []LikwidCollectorMetricConfig `json:"metrics"`
Events map[string]string `json:"events"`
Metrics []LikwidCollectorMetricConfig `json:"metrics"`
}
type LikwidEventsetConfig struct {
internal int
gid C.int
eorder []*C.char
estr *C.char
go_estr string
results map[int]map[string]interface{}
metrics map[int]map[string]float64
}
type LikwidCollectorConfig struct {
@@ -98,28 +67,29 @@ type LikwidCollectorConfig struct {
InvalidToZero bool `json:"invalid_to_zero,omitempty"`
AccessMode string `json:"access_mode,omitempty"`
DaemonPath string `json:"accessdaemon_path,omitempty"`
LibraryPath string `json:"liblikwid_path,omitempty"`
}
type LikwidCollector struct {
metricCollector
cpulist []C.int
cpu2tid map[int]int
sock2tid map[int]int
scopeRespTids map[MetricScope]map[int]int
metrics map[C.int]map[string]int
groups []C.int
config LikwidCollectorConfig
results map[int]map[int]map[string]interface{}
mresults map[int]map[int]map[string]float64
gmresults map[int]map[string]float64
basefreq float64
running bool
cpulist []C.int
cpu2tid map[int]int
sock2tid map[int]int
metrics map[C.int]map[string]int
groups []C.int
config LikwidCollectorConfig
gmresults map[int]map[string]float64
basefreq float64
running bool
initialized bool
likwidGroups map[C.int]LikwidEventsetConfig
lock sync.Mutex
}
type LikwidMetric struct {
name string
search string
scope MetricScope
scope string
group_idx int
}
@@ -131,152 +101,100 @@ func eventsToEventStr(events map[string]string) string {
return strings.Join(elist, ",")
}
func getGranularity(counter, event string) MetricScope {
if strings.HasPrefix(counter, "PMC") || strings.HasPrefix(counter, "FIXC") {
return "cpu"
} else if strings.Contains(counter, "BOX") || strings.Contains(counter, "DEV") {
return "socket"
} else if strings.HasPrefix(counter, "PWR") {
if event == "RAPL_CORE_ENERGY" {
return "cpu"
} else {
return "socket"
func genLikwidEventSet(input LikwidCollectorEventsetConfig) LikwidEventsetConfig {
tmplist := make([]string, 0)
clist := make([]string, 0)
for k := range input.Events {
clist = append(clist, k)
}
sort.Strings(clist)
elist := make([]*C.char, 0)
for _, k := range clist {
v := input.Events[k]
tmplist = append(tmplist, fmt.Sprintf("%s:%s", v, k))
c_counter := C.CString(k)
elist = append(elist, c_counter)
}
estr := strings.Join(tmplist, ",")
res := make(map[int]map[string]interface{})
met := make(map[int]map[string]float64)
for _, i := range topo.CpuList() {
res[i] = make(map[string]interface{})
for k := range input.Events {
res[i][k] = 0.0
}
met[i] = make(map[string]float64)
for _, v := range input.Metrics {
res[i][v.Name] = 0.0
}
}
return "unknown"
return LikwidEventsetConfig{
gid: -1,
eorder: elist,
estr: C.CString(estr),
go_estr: estr,
results: res,
metrics: met,
}
}
func testLikwidMetricFormula(formula string, params []string) bool {
myparams := make(map[string]interface{})
for _, p := range params {
myparams[p] = float64(1.0)
}
_, err := agg.EvalFloat64Condition(formula, myparams)
return err == nil
}
func getBaseFreq() float64 {
files := []string{
"/sys/devices/system/cpu/cpu0/cpufreq/bios_limit",
"/sys/devices/system/cpu/cpu0/cpufreq/base_frequency",
}
var freq float64 = math.NaN()
C.power_init(0)
info := C.get_powerInfo()
if float64(info.baseFrequency) != 0 {
freq = float64(info.baseFrequency) * 1e3
} else {
buffer, err := ioutil.ReadFile("/sys/devices/system/cpu/cpu0/cpufreq/bios_limit")
for _, f := range files {
buffer, err := ioutil.ReadFile(f)
if err == nil {
data := strings.Replace(string(buffer), "\n", "", -1)
x, err := strconv.ParseInt(data, 0, 64)
if err == nil {
freq = float64(x) * 1e3
freq = float64(x) * 1e6
}
}
}
if math.IsNaN(freq) {
C.power_init(0)
info := C.get_powerInfo()
if float64(info.baseFrequency) != 0 {
freq = float64(info.baseFrequency) * 1e6
}
C.power_finalize()
}
return freq
}
func (m *LikwidCollector) initGranularity() {
splitRegex := regexp.MustCompile("[+-/*()]")
for _, evset := range m.config.Eventsets {
evset.granulatity = make(map[string]MetricScope)
for counter, event := range evset.Events {
gran := getGranularity(counter, event)
if gran.Granularity() >= 0 {
evset.granulatity[counter] = gran
}
}
for i, metric := range evset.Metrics {
s := splitRegex.Split(metric.Calc, -1)
gran := MetricScope("cpu")
evset.Metrics[i].granulatity = gran
for _, x := range s {
if _, ok := evset.Events[x]; ok {
if evset.granulatity[x].Granularity() > gran.Granularity() {
gran = evset.granulatity[x]
}
}
}
evset.Metrics[i].granulatity = gran
}
}
for i, metric := range m.config.Metrics {
s := splitRegex.Split(metric.Calc, -1)
gran := MetricScope("cpu")
m.config.Metrics[i].granulatity = gran
for _, x := range s {
for _, evset := range m.config.Eventsets {
for _, m := range evset.Metrics {
if m.Name == x && m.granulatity.Granularity() > gran.Granularity() {
gran = m.granulatity
}
}
}
}
m.config.Metrics[i].granulatity = gran
}
}
type TopoResolveFunc func(cpuid int) int
func (m *LikwidCollector) getResponsiblities() map[MetricScope]map[int]int {
get_cpus := func(scope MetricScope) map[int]int {
var slist []int
var cpu C.int
var input func(index int) string
switch scope {
case "node":
slist = []int{0}
input = func(index int) string { return "N:0" }
case "socket":
input = func(index int) string { return fmt.Sprintf("%s%d:0", scope.Likwid(), index) }
slist = topo.SocketList()
// case "numadomain":
// input = func(index int) string { return fmt.Sprintf("%s%d:0", scope.Likwid(), index) }
// slist = topo.NumaNodeList()
// cclog.Debug(scope, " ", input(0), " ", slist)
// case "die":
// input = func(index int) string { return fmt.Sprintf("%s%d:0", scope.Likwid(), index) }
// slist = topo.DieList()
// case "llc":
// input = fmt.Sprintf("%s%d:0", scope.Likwid(), s)
// slist = topo.LLCacheList()
case "cpu":
input = func(index int) string { return fmt.Sprintf("%d", index) }
slist = topo.CpuList()
case "hwthread":
input = func(index int) string { return fmt.Sprintf("%d", index) }
slist = topo.CpuList()
}
outmap := make(map[int]int)
for _, s := range slist {
t := C.CString(input(s))
clen := C.cpustr_to_cpulist(t, &cpu, 1)
if int(clen) == 1 {
outmap[s] = m.cpu2tid[int(cpu)]
} else {
cclog.Error(fmt.Sprintf("Cannot determine responsible CPU for %s", input(s)))
outmap[s] = -1
}
C.free(unsafe.Pointer(t))
}
return outmap
}
scopes := GetAllMetricScopes()
complete := make(map[MetricScope]map[int]int)
for _, s := range scopes {
complete[s] = get_cpus(s)
}
return complete
}
func (m *LikwidCollector) Init(config json.RawMessage) error {
var ret C.int
m.name = "LikwidCollector"
m.parallel = false
m.initialized = false
m.running = false
m.config.AccessMode = LIKWID_DEF_ACCESSMODE
m.config.LibraryPath = LIKWID_LIB_NAME
if len(config) > 0 {
err := json.Unmarshal(config, &m.config)
if err != nil {
return err
}
}
lib := dl.New(LIKWID_LIB_NAME, LIKWID_LIB_DL_FLAGS)
lib := dl.New(m.config.LibraryPath, LIKWID_LIB_DL_FLAGS)
if lib == nil {
return fmt.Errorf("error instantiating DynamicLibrary for %s", LIKWID_LIB_NAME)
return fmt.Errorf("error instantiating DynamicLibrary for %s", m.config.LibraryPath)
}
err := lib.Open()
if err != nil {
return fmt.Errorf("error opening %s: %v", LIKWID_LIB_NAME, err)
return fmt.Errorf("error opening %s: %v", m.config.LibraryPath, err)
}
if m.config.ForceOverwrite {
@@ -285,202 +203,174 @@ func (m *LikwidCollector) Init(config json.RawMessage) error {
}
m.setup()
m.meta = map[string]string{"source": m.name, "group": "PerfCounter"}
m.meta = map[string]string{"group": "PerfCounter"}
cclog.ComponentDebug(m.name, "Get cpulist and init maps and lists")
cpulist := topo.CpuList()
cpulist := topo.HwthreadList()
m.cpulist = make([]C.int, len(cpulist))
m.cpu2tid = make(map[int]int)
for i, c := range cpulist {
m.cpulist[i] = C.int(c)
m.cpu2tid[c] = i
}
m.results = make(map[int]map[int]map[string]interface{})
m.mresults = make(map[int]map[int]map[string]float64)
m.likwidGroups = make(map[C.int]LikwidEventsetConfig)
// m.results = make(map[int]map[int]map[string]interface{})
// m.mresults = make(map[int]map[int]map[string]float64)
m.gmresults = make(map[int]map[string]float64)
cclog.ComponentDebug(m.name, "initialize LIKWID topology")
ret = C.topology_init()
if ret != 0 {
err := errors.New("failed to initialize LIKWID topology")
cclog.ComponentError(m.name, err.Error())
return err
}
// Determine which counter works at which level. PMC*: cpu, *BOX*: socket, ...
m.initGranularity()
// Generate map for MetricScope -> scope_id (like socket id) -> responsible id (offset in cpulist)
m.scopeRespTids = m.getResponsiblities()
switch m.config.AccessMode {
case "direct":
C.HPMmode(0)
case "accessdaemon":
if len(m.config.DaemonPath) > 0 {
p := os.Getenv("PATH")
os.Setenv("PATH", m.config.DaemonPath+":"+p)
}
C.HPMmode(1)
}
cclog.ComponentDebug(m.name, "initialize LIKWID perfmon module")
ret = C.perfmon_init(C.int(len(m.cpulist)), &m.cpulist[0])
if ret != 0 {
C.topology_finalize()
err := errors.New("failed to initialize LIKWID topology")
cclog.ComponentError(m.name, err.Error())
return err
for _, tid := range m.cpu2tid {
m.gmresults[tid] = make(map[string]float64)
}
// This is for the global metrics computation test
globalParams := make(map[string]interface{})
globalParams["time"] = float64(1.0)
globalParams["inverseClock"] = float64(1.0)
// While adding the events, we test the metrics whether they can be computed at all
for i, evset := range m.config.Eventsets {
estr := eventsToEventStr(evset.Events)
// Generate parameter list for the metric computing test
params := make(map[string]interface{})
params["time"] = float64(1.0)
params["inverseClock"] = float64(1.0)
for counter := range evset.Events {
params[counter] = float64(1.0)
}
for _, metric := range evset.Metrics {
// Try to evaluate the metric
_, err := agg.EvalFloat64Condition(metric.Calc, params)
if err != nil {
cclog.ComponentError(m.name, "Calculation for metric", metric.Name, "failed:", err.Error())
continue
totalMetrics := 0
// Generate parameter list for the metric computing test
params := make([]string, 0)
params = append(params, "time", "inverseClock")
// Generate parameter list for the global metric computing test
globalParams := make([]string, 0)
globalParams = append(globalParams, "time", "inverseClock")
// We test the eventset metrics whether they can be computed at all
for _, evset := range m.config.Eventsets {
if len(evset.Events) > 0 {
params = params[:2]
for counter := range evset.Events {
params = append(params, counter)
}
// If the metric is not in the parameter list for the global metrics, add it
if _, ok := globalParams[metric.Name]; !ok {
globalParams[metric.Name] = float64(1.0)
}
}
// Now we add the list of events to likwid
cstr := C.CString(estr)
gid := C.perfmon_addEventSet(cstr)
if gid >= 0 {
m.groups = append(m.groups, gid)
}
C.free(unsafe.Pointer(cstr))
m.results[i] = make(map[int]map[string]interface{})
m.mresults[i] = make(map[int]map[string]float64)
for tid := range m.cpulist {
m.results[i][tid] = make(map[string]interface{})
m.mresults[i][tid] = make(map[string]float64)
if i == 0 {
m.gmresults[tid] = make(map[string]float64)
for _, metric := range evset.Metrics {
// Try to evaluate the metric
if testLikwidMetricFormula(metric.Calc, params) {
// Add the computable metric to the parameter list for the global metrics
globalParams = append(globalParams, metric.Name)
totalMetrics++
} else {
metric.Calc = ""
}
}
} else {
cclog.ComponentError(m.name, "Invalid Likwid eventset config, no events given")
continue
}
}
for _, metric := range m.config.Metrics {
// Try to evaluate the global metric
_, err := agg.EvalFloat64Condition(metric.Calc, globalParams)
if err != nil {
cclog.ComponentError(m.name, "Calculation for metric", metric.Name, "failed:", err.Error())
continue
if !testLikwidMetricFormula(metric.Calc, globalParams) {
cclog.ComponentError(m.name, "Calculation for metric", metric.Name, "failed")
metric.Calc = ""
} else {
totalMetrics++
}
}
// If no event set could be added, shut down LikwidCollector
if len(m.groups) == 0 {
C.perfmon_finalize()
C.topology_finalize()
err := errors.New("no LIKWID performance group initialized")
if totalMetrics == 0 {
err := errors.New("no LIKWID eventset or metric usable")
cclog.ComponentError(m.name, err.Error())
return err
}
m.basefreq = getBaseFreq()
cclog.ComponentDebug(m.name, "BaseFreq", m.basefreq)
m.init = true
return nil
}
// take a measurement for 'interval' seconds of event set index 'group'
func (m *LikwidCollector) takeMeasurement(group int, interval time.Duration) error {
func (m *LikwidCollector) takeMeasurement(evset LikwidEventsetConfig, interval time.Duration) (bool, error) {
var ret C.int
gid := m.groups[group]
ret = C.perfmon_setupCounters(gid)
if ret != 0 {
gctr := C.GoString(C.perfmon_getGroupName(gid))
err := fmt.Errorf("failed to setup performance group %d (%s)", gid, gctr)
return err
m.lock.Lock()
if m.initialized {
ret = C.perfmon_setupCounters(evset.gid)
if ret != 0 {
var err error = nil
var skip bool = false
if ret == -37 {
skip = true
} else {
err = fmt.Errorf("failed to setup performance group %d", evset.gid)
}
m.lock.Unlock()
return skip, err
}
ret = C.perfmon_startCounters()
if ret != 0 {
var err error = nil
var skip bool = false
if ret == -37 {
skip = true
} else {
err = fmt.Errorf("failed to setup performance group %d", evset.gid)
}
m.lock.Unlock()
return skip, err
}
m.running = true
time.Sleep(interval)
m.running = false
ret = C.perfmon_stopCounters()
if ret != 0 {
var err error = nil
var skip bool = false
if ret == -37 {
skip = true
} else {
err = fmt.Errorf("failed to setup performance group %d", evset.gid)
}
m.lock.Unlock()
return skip, err
}
}
ret = C.perfmon_startCounters()
if ret != 0 {
gctr := C.GoString(C.perfmon_getGroupName(gid))
err := fmt.Errorf("failed to start performance group %d (%s)", gid, gctr)
return err
}
m.running = true
time.Sleep(interval)
m.running = false
ret = C.perfmon_stopCounters()
if ret != 0 {
gctr := C.GoString(C.perfmon_getGroupName(gid))
err := fmt.Errorf("failed to stop performance group %d (%s)", gid, gctr)
return err
}
return nil
m.lock.Unlock()
return false, nil
}
// Get all measurement results for an event set, derive the metric values out of the measurement results and send it
func (m *LikwidCollector) calcEventsetMetrics(group int, interval time.Duration, output chan lp.CCMetric) error {
var eidx C.int
evset := m.config.Eventsets[group]
gid := m.groups[group]
func (m *LikwidCollector) calcEventsetMetrics(evset LikwidEventsetConfig, interval time.Duration, output chan lp.CCMetric) error {
invClock := float64(1.0 / m.basefreq)
// Go over events and get the results
for eidx = 0; int(eidx) < len(evset.Events); eidx++ {
ctr := C.perfmon_getCounterName(gid, eidx)
ev := C.perfmon_getEventName(gid, eidx)
gctr := C.GoString(ctr)
gev := C.GoString(ev)
// MetricScope for the counter (and if needed the event)
scope := getGranularity(gctr, gev)
// Get the map scope-id -> tids
// This way we read less counters like only the responsible hardware thread for a socket
scopemap := m.scopeRespTids[scope]
for _, tid := range scopemap {
if tid >= 0 {
m.results[group][tid]["time"] = interval.Seconds()
m.results[group][tid]["inverseClock"] = invClock
res := C.perfmon_getLastResult(gid, eidx, C.int(tid))
m.results[group][tid][gctr] = float64(res)
for eidx, counter := range evset.eorder {
gctr := C.GoString(counter)
for _, tid := range m.cpu2tid {
res := C.perfmon_getLastResult(evset.gid, C.int(eidx), C.int(tid))
fres := float64(res)
if m.config.InvalidToZero && (math.IsNaN(fres) || math.IsInf(fres, 0)) {
fres = 0.0
}
evset.results[tid][gctr] = fres
evset.results[tid]["time"] = interval.Seconds()
evset.results[tid]["inverseClock"] = invClock
}
}
// Go over the event set metrics, derive the value out of the event:counter values and send it
for _, metric := range evset.Metrics {
for _, metric := range m.config.Eventsets[evset.internal].Metrics {
// The metric scope is determined in the Init() function
// Get the map scope-id -> tids
scopemap := m.scopeRespTids[metric.Scope]
scopemap := m.cpu2tid
if metric.Type == "socket" {
scopemap = m.sock2tid
}
for domain, tid := range scopemap {
if tid >= 0 {
value, err := agg.EvalFloat64Condition(metric.Calc, m.results[group][tid])
if tid >= 0 && len(metric.Calc) > 0 {
value, err := agg.EvalFloat64Condition(metric.Calc, evset.results[tid])
if err != nil {
cclog.ComponentError(m.name, "Calculation for metric", metric.Name, "failed:", err.Error())
continue
}
m.mresults[group][tid][metric.Name] = value
if m.config.InvalidToZero && math.IsNaN(value) {
value = 0.0
}
if m.config.InvalidToZero && math.IsInf(value, 0) {
if m.config.InvalidToZero && (math.IsNaN(value) || math.IsInf(value, 0)) {
value = 0.0
}
evset.metrics[tid][metric.Name] = value
// Now we have the result, send it with the proper tags
if !math.IsNaN(value) {
if metric.Publish {
tags := map[string]string{"type": metric.Scope.String()}
if metric.Scope != "node" {
tags["type-id"] = fmt.Sprintf("%d", domain)
}
fields := map[string]interface{}{"value": value}
y, err := lp.New(metric.Name, tags, m.meta, fields, time.Now())
y, err := lp.New(metric.Name, map[string]string{"type": metric.Type}, m.meta, fields, time.Now())
if err == nil {
if metric.Type != "node" {
y.AddTag("type-id", fmt.Sprintf("%d", domain))
}
if len(metric.Unit) > 0 {
y.AddMeta("unit", metric.Unit)
}
output <- y
}
}
@@ -495,13 +385,16 @@ func (m *LikwidCollector) calcEventsetMetrics(group int, interval time.Duration,
// Go over the global metrics, derive the value out of the event sets' metric values and send it
func (m *LikwidCollector) calcGlobalMetrics(interval time.Duration, output chan lp.CCMetric) error {
for _, metric := range m.config.Metrics {
scopemap := m.scopeRespTids[metric.Scope]
scopemap := m.cpu2tid
if metric.Type == "socket" {
scopemap = m.sock2tid
}
for domain, tid := range scopemap {
if tid >= 0 {
// Here we generate parameter list
params := make(map[string]interface{})
for j := range m.groups {
for mname, mres := range m.mresults[j][tid] {
for _, evset := range m.likwidGroups {
for mname, mres := range evset.metrics[tid] {
params[mname] = mres
}
}
@@ -509,25 +402,25 @@ func (m *LikwidCollector) calcGlobalMetrics(interval time.Duration, output chan
value, err := agg.EvalFloat64Condition(metric.Calc, params)
if err != nil {
cclog.ComponentError(m.name, "Calculation for metric", metric.Name, "failed:", err.Error())
continue
value = 0.0
}
if m.config.InvalidToZero && (math.IsNaN(value) || math.IsInf(value, 0)) {
value = 0.0
}
m.gmresults[tid][metric.Name] = value
if m.config.InvalidToZero && math.IsNaN(value) {
value = 0.0
}
if m.config.InvalidToZero && math.IsInf(value, 0) {
value = 0.0
}
// Now we have the result, send it with the proper tags
if !math.IsNaN(value) {
if metric.Publish {
tags := map[string]string{"type": metric.Scope.String()}
if metric.Scope != "node" {
tags["type-id"] = fmt.Sprintf("%d", domain)
}
tags := map[string]string{"type": metric.Type}
fields := map[string]interface{}{"value": value}
y, err := lp.New(metric.Name, tags, m.meta, fields, time.Now())
if err == nil {
if metric.Type != "node" {
y.AddTag("type-id", fmt.Sprintf("%d", domain))
}
if len(metric.Unit) > 0 {
y.AddMeta("unit", metric.Unit)
}
output <- y
}
}
@@ -538,38 +431,163 @@ func (m *LikwidCollector) calcGlobalMetrics(interval time.Duration, output chan
return nil
}
func (m *LikwidCollector) LateInit() error {
var ret C.int
if m.initialized {
return nil
}
switch m.config.AccessMode {
case "direct":
C.HPMmode(0)
case "accessdaemon":
if len(m.config.DaemonPath) > 0 {
p := os.Getenv("PATH")
os.Setenv("PATH", m.config.DaemonPath+":"+p)
}
C.HPMmode(1)
}
cclog.ComponentDebug(m.name, "initialize LIKWID topology")
ret = C.topology_init()
if ret != 0 {
err := errors.New("failed to initialize LIKWID topology")
cclog.ComponentError(m.name, err.Error())
return err
}
m.sock2tid = make(map[int]int)
tmp := make([]C.int, 1)
for _, sid := range topo.SocketList() {
cstr := C.CString(fmt.Sprintf("S%d:0", sid))
ret = C.cpustr_to_cpulist(cstr, &tmp[0], 1)
if ret > 0 {
m.sock2tid[sid] = m.cpu2tid[int(tmp[0])]
}
C.free(unsafe.Pointer(cstr))
}
m.basefreq = getBaseFreq()
cclog.ComponentDebug(m.name, "BaseFreq", m.basefreq)
cclog.ComponentDebug(m.name, "initialize LIKWID perfmon module")
ret = C.perfmon_init(C.int(len(m.cpulist)), &m.cpulist[0])
if ret != 0 {
var err error = nil
C.topology_finalize()
if ret != -22 {
err = errors.New("failed to initialize LIKWID perfmon")
cclog.ComponentError(m.name, err.Error())
} else {
err = errors.New("access to LIKWID perfmon locked")
}
return err
}
// While adding the events, we test the metrics whether they can be computed at all
for i, evset := range m.config.Eventsets {
var gid C.int
if len(evset.Events) > 0 {
skip := false
likwidGroup := genLikwidEventSet(evset)
for _, g := range m.likwidGroups {
if likwidGroup.go_estr == g.go_estr {
skip = true
break
}
}
if skip {
continue
}
// Now we add the list of events to likwid
gid = C.perfmon_addEventSet(likwidGroup.estr)
if gid >= 0 {
likwidGroup.gid = gid
likwidGroup.internal = i
m.likwidGroups[gid] = likwidGroup
}
} else {
cclog.ComponentError(m.name, "Invalid Likwid eventset config, no events given")
continue
}
}
// If no event set could be added, shut down LikwidCollector
if len(m.likwidGroups) == 0 {
C.perfmon_finalize()
C.topology_finalize()
err := errors.New("no LIKWID performance group initialized")
cclog.ComponentError(m.name, err.Error())
return err
}
sigchan := make(chan os.Signal, 1)
signal.Notify(sigchan, syscall.SIGCHLD)
signal.Notify(sigchan, os.Interrupt)
go func() {
<-sigchan
signal.Stop(sigchan)
m.initialized = false
}()
m.initialized = true
return nil
}
// main read function taking multiple measurement rounds, each 'interval' seconds long
func (m *LikwidCollector) Read(interval time.Duration, output chan lp.CCMetric) {
var skip bool = false
var err error
if !m.init {
return
}
for i := range m.groups {
// measure event set 'i' for 'interval' seconds
err := m.takeMeasurement(i, interval)
if !m.initialized {
m.lock.Lock()
err = m.LateInit()
if err != nil {
cclog.ComponentError(m.name, err.Error())
m.lock.Unlock()
return
}
// read measurements and derive event set metrics
m.calcEventsetMetrics(i, interval, output)
m.initialized = true
m.lock.Unlock()
}
if m.initialized && !skip {
for _, evset := range m.likwidGroups {
if !skip {
// measure event set 'i' for 'interval' seconds
skip, err = m.takeMeasurement(evset, interval)
if err != nil {
cclog.ComponentError(m.name, err.Error())
return
}
}
if !skip {
// read measurements and derive event set metrics
m.calcEventsetMetrics(evset, interval, output)
}
}
if !skip {
// use the event set metrics to derive the global metrics
m.calcGlobalMetrics(interval, output)
}
}
// use the event set metrics to derive the global metrics
m.calcGlobalMetrics(interval, output)
}
func (m *LikwidCollector) Close() {
if m.init {
cclog.ComponentDebug(m.name, "Closing ...")
m.init = false
if m.running {
cclog.ComponentDebug(m.name, "Stopping counters")
C.perfmon_stopCounters()
cclog.ComponentDebug(m.name, "Closing ...")
m.lock.Lock()
if m.initialized {
cclog.ComponentDebug(m.name, "Finalize LIKWID perfmon module")
C.perfmon_finalize()
m.initialized = false
}
cclog.ComponentDebug(m.name, "Finalize LIKWID perfmon module")
C.perfmon_finalize()
m.lock.Unlock()
cclog.ComponentDebug(m.name, "Finalize LIKWID topology module")
C.topology_finalize()
cclog.ComponentDebug(m.name, "Closing done")
}
}

View File

@@ -3,29 +3,63 @@
The `likwid` collector is probably the most complicated collector. The LIKWID library is included as static library with *direct* access mode. The *direct* access mode is suitable if the daemon is executed by a root user. The static library does not contain the performance groups, so all information needs to be provided in the configuration.
The `likwid` configuration consists of two parts, the "eventsets" and "globalmetrics":
- An event set list itself has two parts, the "events" and a set of derivable "metrics". Each of the "events" is a counter:event pair in LIKWID's syntax. The "metrics" are a list of formulas to derive the metric value from the measurements of the "events". Each metric has a name, the formula, a scope and a publish flag. Counter names can be used like variables in the formulas, so `PMC0+PMC1` sums the measurements for the both events configured in the counters `PMC0` and `PMC1`. The scope tells the Collector whether it is a metric for each hardware thread (`cpu`) or each CPU socket (`socket`). The last one is the publishing flag. It tells the collector whether a metric should be sent to the router.
- The global metrics are metrics which require data from all event set measurements to be derived. The inputs are the metrics in the event sets. Similar to the metrics in the event sets, the global metrics are defined by a name, a formula, a scope and a publish flag. See event set metrics for details. The only difference is that there is no access to the raw event measurements anymore but only to the metrics. So, the idea is to derive a metric in the "eventsets" section and reuse it in the "globalmetrics" part. If you need a metric only for deriving the global metrics, disable forwarding of the event set metrics. **Be aware** that the combination might be misleading because the "behavior" of a metric changes over time and the multiple measurements might count different computing phases.
```json
"likwid": {
"force_overwrite" : false,
"invalid_to_zero" : false,
"eventsets": [
{
"events" : {
"COUNTER0": "EVENT0",
"COUNTER1": "EVENT1",
},
"metrics" : [
{
"name": "sum_01",
"calc": "COUNTER0 + COUNTER1",
"publish": false,
"unit": "myunit",
"type": "hwthread"
}
]
}
]
"globalmetrics" : [
{
"name": "global_sum",
"calc": "sum_01",
"publish": true,
"unit": "myunit",
"type": "hwthread"
}
]
}
```
The `likwid` configuration consists of two parts, the `eventsets` and `globalmetrics`:
- An event set list itself has two parts, the `events` and a set of derivable `metrics`. Each of the `events` is a `counter:event` pair in LIKWID's syntax. The `metrics` are a list of formulas to derive the metric value from the measurements of the `events`' values. Each metric has a name, the formula, a type and a publish flag. There is an optional `unit` field. Counter names can be used like variables in the formulas, so `PMC0+PMC1` sums the measurements for the both events configured in the counters `PMC0` and `PMC1`. You can optionally use `time` for the measurement time and `inverseClock` for `1.0/baseCpuFrequency`. The type tells the LikwidCollector whether it is a metric for each hardware thread (`cpu`) or each CPU socket (`socket`). You may specify a unit for the metric with `unit`. The last one is the publishing flag. It tells the LikwidCollector whether a metric should be sent to the router or is only used internally to compute a global metric.
- The `globalmetrics` are metrics which require data from multiple event set measurements to be derived. The inputs are the metrics in the event sets. Similar to the metrics in the event sets, the global metrics are defined by a name, a formula, a scope and a publish flag. See event set metrics for details. The only difference is that there is no access to the raw event measurements anymore but only to the metrics. Also `time` and `inverseClock` cannot be used anymore. So, the idea is to derive a metric in the `eventsets` section and reuse it in the `globalmetrics` part. If you need a metric only for deriving the global metrics, disable forwarding of the event set metrics (`"publish": false`). **Be aware** that the combination might be misleading because the "behavior" of a metric changes over time and the multiple measurements might count different computing phases. Similar to the metrics in the eventset, you can specify a metric unit with the `unit` field.
Additional options:
- `access_mode` : Method to use for hardware performance monitoring (`direct` access as root user, `accessdaemon` for the daemon mode)
- `accessdaemon_path`: Folder with the access daemon `likwid-accessD`, commonly `$LIKWID_INSTALL_LOC/sbin`
- `force_overwrite`: Same as setting `LIKWID_FORCE=1`. In case counters are already in-use, LIKWID overwrites their configuration to do its measurements
- `invalid_to_zero`: In some cases, the calculations result in `NaN` or `Inf`. With this option, all `NaN` and `Inf` values are replaces with `0.0`.
- `invalid_to_zero`: In some cases, the calculations result in `NaN` or `Inf`. With this option, all `NaN` and `Inf` values are replaces with `0.0`. See below in [seperate section](./likwidMetric.md#invalid_to_zero-option)
- `access_mode`: Specify LIKWID access mode: `direct` for direct register access as root user or `accessdaemon`. The access mode `perf_event` is current untested.
- `accessdaemon_path`: Folder of the accessDaemon `likwid-accessD` (like `/usr/local/sbin`)
- `liblikwid_path`: Location of `liblikwid.so` including file name like `/usr/local/lib/liblikwid.so`
### Available metric scopes
Hardware performance counters are scattered all over the system nowadays. A counter coveres a specific part of the system. While there are hardware thread specific counter for CPU cycles, instructions and so on, some others are specific for a whole CPU socket/package. To address that, the collector provides the specification of a 'scope' for each metric.
Hardware performance counters are scattered all over the system nowadays. A counter coveres a specific part of the system. While there are hardware thread specific counter for CPU cycles, instructions and so on, some others are specific for a whole CPU socket/package. To address that, the LikwidCollector provides the specification of a `type` for each metric.
- `cpu` : One metric per CPU hardware thread with the tags `"type" : "cpu"` and `"type-id" : "$cpu_id"`
- `hwthread` : One metric per CPU hardware thread with the tags `"type" : "hwthread"` and `"type-id" : "$hwthread_id"`
- `socket` : One metric per CPU socket/package with the tags `"type" : "socket"` and `"type-id" : "$socket_id"`
**Note:** You cannot specify `socket` scope for a metric that is measured at `cpu` scope, so some kind of expert knowledge or lookup work in the [Likwid Wiki](https://github.com/RRZE-HPC/likwid/wiki) is required. Get the scope of each counter from the *Architecture* pages and as soon as one counter in a metric is socket-specific, the whole metric is socket-specific.
**Note:** You cannot specify `socket` scope for a metric that is measured at `hwthread` scope, so some kind of expert knowledge or lookup work in the [Likwid Wiki](https://github.com/RRZE-HPC/likwid/wiki) is required. Get the scope of each counter from the *Architecture* pages and as soon as one counter in a metric is socket-specific, the whole metric is socket-specific.
As a guideline:
- All counters `FIXCx`, `PMCy` and `TMAz` have the scope `cpu`
- All counters `FIXCx`, `PMCy` and `TMAz` have the scope `hwthread`
- All counters names containing `BOX` have the scope `socket`
- All `PWRx` counters have scope `socket`, except `"PWR1" : "RAPL_CORE_ENERGY"` has `cpu` scope
- All `PWRx` counters have scope `socket`, except `"PWR1" : "RAPL_CORE_ENERGY"` has `hwthread` scope
- All `DFCx` counters have scope `socket`
### Help with the configuration
@@ -47,6 +81,7 @@ $ scripts/likwid_perfgroup_to_cc_config.py ICX MEM_DP
{
"events": {
"FIXC0": "INSTR_RETIRED_ANY",
"FIXC1": "CPU_CLK_UNHALTED_CORE",
"..." : "..."
},
"metrics" : [
@@ -54,6 +89,7 @@ $ scripts/likwid_perfgroup_to_cc_config.py ICX MEM_DP
"calc": "time",
"name": "Runtime (RDTSC) [s]",
"publish": true,
"unit": "seconds"
"scope": "hwthread"
},
{
@@ -71,21 +107,28 @@ LIKWID checks the file `/var/run/likwid.lock` before performing any interfering
Before (SLURM prolog, ...)
```
$ chwon $JOBUSER /var/run/likwid.lock
$ chown $JOBUSER /var/run/likwid.lock
```
After (SLURM epilog, ...)
```
$ chwon $CCUSER /var/run/likwid.lock
$ chown $CCUSER /var/run/likwid.lock
```
### `invalid_to_zero` option
In some cases LIKWID returns `0.0` for some events that are further used in processing and maybe used as divisor in a calculation. After evaluation of a metric, the result might be `NaN` or `+-Inf`. These resulting metrics are commonly not created and forwarded to the router because the [InfluxDB line protocol](https://docs.influxdata.com/influxdb/cloud/reference/syntax/line-protocol/#float) does not support these special floating-point values. If you want to have them sent, this option forces these metric values to be `0.0` instead.
One might think this does not happen often but often used metrics in the world of performance engineering like Instructions-per-Cycle (IPC) or more frequently the actual CPU clock are derived with events like `CPU_CLK_UNHALTED_CORE` (Intel) which do not increment in halted state (as the name implies). In there are different power management systems in a chip which can cause a hardware thread to go in such a state. Moreover, if no cycles are executed by the core, also many other events are not incremented as well (like `INSTR_RETIRED_ANY` for retired instructions and part of IPC).
### Example configuration
#### AMD Zen3
```json
"likwid": {
"force_overwrite" : false,
"nan_to_zero" : false,
"invalid_to_zero" : false,
"eventsets": [
{
"events": {
@@ -104,25 +147,28 @@ $ chwon $CCUSER /var/run/likwid.lock
{
"name": "ipc",
"calc": "PMC0/PMC1",
"scope": "cpu",
"type": "hwthread",
"publish": true
},
{
"name": "flops_any",
"calc": "0.000001*PMC2/time",
"scope": "cpu",
"unit": "MFlops/s",
"type": "hwthread",
"publish": true
},
{
"name": "clock_mhz",
"name": "clock",
"calc": "0.000001*(FIXC1/FIXC2)/inverseClock",
"scope": "cpu",
"type": "hwthread",
"unit": "MHz",
"publish": true
},
{
"name": "mem1",
"calc": "0.000001*(DFC0+DFC1+DFC2+DFC3)*64.0/time",
"scope": "socket",
"unit": "Mbyte/s",
"type": "socket",
"publish": false
}
]
@@ -140,19 +186,22 @@ $ chwon $CCUSER /var/run/likwid.lock
{
"name": "pwr_core",
"calc": "PWR0/time",
"scope": "socket",
"unit": "Watt"
"type": "socket",
"publish": true
},
{
"name": "pwr_pkg",
"calc": "PWR1/time",
"scope": "socket",
"type": "socket",
"unit": "Watt"
"publish": true
},
{
"name": "mem2",
"calc": "0.000001*(DFC0+DFC1+DFC2+DFC3)*64.0/time",
"scope": "socket",
"unit": "Mbyte/s",
"type": "socket",
"publish": false
}
]
@@ -162,7 +211,8 @@ $ chwon $CCUSER /var/run/likwid.lock
{
"name": "mem_bw",
"calc": "mem1+mem2",
"scope": "socket",
"type": "socket",
"unit": "Mbyte/s",
"publish": true
}
]
@@ -171,7 +221,7 @@ $ chwon $CCUSER /var/run/likwid.lock
### How to get the eventsets and metrics from LIKWID
The `likwid` collector reads hardware performance counters at a **cpu** and **socket** level. The configuration looks quite complicated but it is basically copy&paste from [LIKWID's performance groups](https://github.com/RRZE-HPC/likwid/tree/master/groups). The collector made multiple iterations and tried to use the performance groups but it lacked flexibility. The current way of configuration provides most flexibility.
The `likwid` collector reads hardware performance counters at a **hwthread** and **socket** level. The configuration looks quite complicated but it is basically copy&paste from [LIKWID's performance groups](https://github.com/RRZE-HPC/likwid/tree/master/groups). The collector made multiple iterations and tried to use the performance groups but it lacked flexibility. The current way of configuration provides most flexibility.
The logic is as following: There are multiple eventsets, each consisting of a list of counters+events and a list of metrics. If you compare a common performance group with the example setting above, there is not much difference:
```
@@ -192,9 +242,10 @@ METRICS -> "metrics": [
IPC PMC0/PMC1 -> {
-> "name" : "IPC",
-> "calc" : "PMC0/PMC1",
-> "scope": "cpu",
-> "scope": "hwthread",
-> "publish": true
-> }
-> ]
```
The script `scripts/likwid_perfgroup_to_cc_config.py` might help you.

View File

@@ -36,6 +36,7 @@ type LoadavgCollector struct {
func (m *LoadavgCollector) Init(config json.RawMessage) error {
m.name = "LoadavgCollector"
m.parallel = true
m.setup()
if len(config) > 0 {
err := json.Unmarshal(config, &m.config)

View File

@@ -19,20 +19,31 @@ const LCTL_CMD = `lctl`
const LCTL_OPTION = `get_param`
type LustreCollectorConfig struct {
LCtlCommand string `json:"lctl_command"`
ExcludeMetrics []string `json:"exclude_metrics"`
SendAllMetrics bool `json:"send_all_metrics"`
Sudo bool `json:"use_sudo"`
LCtlCommand string `json:"lctl_command,omitempty"`
ExcludeMetrics []string `json:"exclude_metrics,omitempty"`
Sudo bool `json:"use_sudo,omitempty"`
SendAbsoluteValues bool `json:"send_abs_values,omitempty"`
SendDerivedValues bool `json:"send_derived_values,omitempty"`
SendDiffValues bool `json:"send_diff_values,omitempty"`
}
type LustreMetricDefinition struct {
name string
lineprefix string
lineoffset int
unit string
calc string
}
type LustreCollector struct {
metricCollector
tags map[string]string
matches map[string]map[string]int
stats map[string]map[string]int64
config LustreCollectorConfig
lctl string
sudoCmd string
tags map[string]string
config LustreCollectorConfig
lctl string
sudoCmd string
lastTimestamp time.Time // Store time stamp of last tick to derive bandwidths
definitions []LustreMetricDefinition // Combined list without excluded metrics
stats map[string]map[string]int64 // Data for last value per device and metric
}
func (m *LustreCollector) getDeviceDataCommand(device string) []string {
@@ -75,6 +86,16 @@ func (m *LustreCollector) getDevices() []string {
return devices
}
func getMetricData(lines []string, prefix string, offset int) (int64, error) {
for _, line := range lines {
if strings.HasPrefix(line, prefix) {
lf := strings.Fields(line)
return strconv.ParseInt(lf[offset], 0, 64)
}
}
return 0, errors.New("no such line in data")
}
// //Version reading the stats data of a device from sysfs
// func (m *LustreCollector) getDeviceDataSysfs(device string) []string {
// llitedir := filepath.Join(LUSTRE_SYSFS, "llite")
@@ -87,9 +108,187 @@ func (m *LustreCollector) getDevices() []string {
// return strings.Split(string(buffer), "\n")
// }
var LustreAbsMetrics = []LustreMetricDefinition{
{
name: "lustre_read_requests",
lineprefix: "read_bytes",
lineoffset: 1,
unit: "requests",
calc: "none",
},
{
name: "lustre_write_requests",
lineprefix: "write_bytes",
lineoffset: 1,
unit: "requests",
calc: "none",
},
{
name: "lustre_read_bytes",
lineprefix: "read_bytes",
lineoffset: 6,
unit: "bytes",
calc: "none",
},
{
name: "lustre_write_bytes",
lineprefix: "write_bytes",
lineoffset: 6,
unit: "bytes",
calc: "none",
},
{
name: "lustre_open",
lineprefix: "open",
lineoffset: 1,
unit: "",
calc: "none",
},
{
name: "lustre_close",
lineprefix: "close",
lineoffset: 1,
unit: "",
calc: "none",
},
{
name: "lustre_setattr",
lineprefix: "setattr",
lineoffset: 1,
unit: "",
calc: "none",
},
{
name: "lustre_getattr",
lineprefix: "getattr",
lineoffset: 1,
unit: "",
calc: "none",
},
{
name: "lustre_statfs",
lineprefix: "statfs",
lineoffset: 1,
unit: "",
calc: "none",
},
{
name: "lustre_inode_permission",
lineprefix: "inode_permission",
lineoffset: 1,
unit: "",
calc: "none",
},
}
var LustreDiffMetrics = []LustreMetricDefinition{
{
name: "lustre_read_requests_diff",
lineprefix: "read_bytes",
lineoffset: 1,
unit: "requests",
calc: "difference",
},
{
name: "lustre_write_requests_diff",
lineprefix: "write_bytes",
lineoffset: 1,
unit: "requests",
calc: "difference",
},
{
name: "lustre_read_bytes_diff",
lineprefix: "read_bytes",
lineoffset: 6,
unit: "bytes",
calc: "difference",
},
{
name: "lustre_write_bytes_diff",
lineprefix: "write_bytes",
lineoffset: 6,
unit: "bytes",
calc: "difference",
},
{
name: "lustre_open_diff",
lineprefix: "open",
lineoffset: 1,
unit: "",
calc: "difference",
},
{
name: "lustre_close_diff",
lineprefix: "close",
lineoffset: 1,
unit: "",
calc: "difference",
},
{
name: "lustre_setattr_diff",
lineprefix: "setattr",
lineoffset: 1,
unit: "",
calc: "difference",
},
{
name: "lustre_getattr_diff",
lineprefix: "getattr",
lineoffset: 1,
unit: "",
calc: "difference",
},
{
name: "lustre_statfs_diff",
lineprefix: "statfs",
lineoffset: 1,
unit: "",
calc: "difference",
},
{
name: "lustre_inode_permission_diff",
lineprefix: "inode_permission",
lineoffset: 1,
unit: "",
calc: "difference",
},
}
var LustreDeriveMetrics = []LustreMetricDefinition{
{
name: "lustre_read_requests_rate",
lineprefix: "read_bytes",
lineoffset: 1,
unit: "requests/sec",
calc: "derivative",
},
{
name: "lustre_write_requests_rate",
lineprefix: "write_bytes",
lineoffset: 1,
unit: "requests/sec",
calc: "derivative",
},
{
name: "lustre_read_bw",
lineprefix: "read_bytes",
lineoffset: 6,
unit: "bytes/sec",
calc: "derivative",
},
{
name: "lustre_write_bw",
lineprefix: "write_bytes",
lineoffset: 6,
unit: "bytes/sec",
calc: "derivative",
},
}
func (m *LustreCollector) Init(config json.RawMessage) error {
var err error
m.name = "LustreCollector"
m.parallel = true
if len(config) > 0 {
err = json.Unmarshal(config, &m.config)
if err != nil {
@@ -99,42 +298,28 @@ func (m *LustreCollector) Init(config json.RawMessage) error {
m.setup()
m.tags = map[string]string{"type": "node"}
m.meta = map[string]string{"source": m.name, "group": "Lustre"}
defmatches := map[string]map[string]int{
"read_bytes": {"lustre_read_bytes": 6, "lustre_read_requests": 1},
"write_bytes": {"lustre_write_bytes": 6, "lustre_write_requests": 1},
"open": {"lustre_open": 1},
"close": {"lustre_close": 1},
"setattr": {"lustre_setattr": 1},
"getattr": {"lustre_getattr": 1},
"statfs": {"lustre_statfs": 1},
"inode_permission": {"lustre_inode_permission": 1}}
// Lustre file system statistics can only be queried by user root
user, err := user.Current()
if err != nil {
cclog.ComponentError(m.name, "Failed to get current user:", err.Error())
return err
}
if user.Uid != "0" {
cclog.ComponentError(m.name, "Lustre file system statistics can only be queried by user root")
return err
// or with password-less sudo
if !m.config.Sudo {
user, err := user.Current()
if err != nil {
cclog.ComponentError(m.name, "Failed to get current user:", err.Error())
return err
}
if user.Uid != "0" {
cclog.ComponentError(m.name, "Lustre file system statistics can only be queried by user root")
return err
}
} else {
p, err := exec.LookPath("sudo")
if err != nil {
cclog.ComponentError(m.name, "Cannot find 'sudo'")
return err
}
m.sudoCmd = p
}
m.matches = make(map[string]map[string]int)
for lineprefix, names := range defmatches {
for metricname, offset := range names {
_, skip := stringArrayContains(m.config.ExcludeMetrics, metricname)
if skip {
continue
}
if _, prefixExist := m.matches[lineprefix]; !prefixExist {
m.matches[lineprefix] = make(map[string]int)
}
if _, metricExist := m.matches[lineprefix][metricname]; !metricExist {
m.matches[lineprefix][metricname] = offset
}
}
}
p, err := exec.LookPath(m.config.LCtlCommand)
if err != nil {
p, err = exec.LookPath(LCTL_CMD)
@@ -143,26 +328,51 @@ func (m *LustreCollector) Init(config json.RawMessage) error {
}
}
m.lctl = p
if m.config.Sudo {
p, err := exec.LookPath("sudo")
if err != nil {
m.sudoCmd = p
m.definitions = []LustreMetricDefinition{}
if m.config.SendAbsoluteValues {
for _, def := range LustreAbsMetrics {
if _, skip := stringArrayContains(m.config.ExcludeMetrics, def.name); !skip {
m.definitions = append(m.definitions, def)
}
}
}
if m.config.SendDiffValues {
for _, def := range LustreDiffMetrics {
if _, skip := stringArrayContains(m.config.ExcludeMetrics, def.name); !skip {
m.definitions = append(m.definitions, def)
}
}
}
if m.config.SendDerivedValues {
for _, def := range LustreDeriveMetrics {
if _, skip := stringArrayContains(m.config.ExcludeMetrics, def.name); !skip {
m.definitions = append(m.definitions, def)
}
}
}
if len(m.definitions) == 0 {
return errors.New("no metrics to collect")
}
devices := m.getDevices()
if len(devices) == 0 {
return errors.New("no metrics to collect")
return errors.New("no Lustre devices found")
}
m.stats = make(map[string]map[string]int64)
for _, d := range devices {
m.stats[d] = make(map[string]int64)
for _, names := range m.matches {
for metricname := range names {
m.stats[d][metricname] = 0
data := m.getDeviceDataCommand(d)
for _, def := range m.definitions {
x, err := getMetricData(data, def.lineprefix, def.lineoffset)
if err == nil {
m.stats[d][def.name] = x
} else {
m.stats[d][def.name] = 0
}
}
}
m.lastTimestamp = time.Now()
m.init = true
return nil
}
@@ -171,54 +381,49 @@ func (m *LustreCollector) Read(interval time.Duration, output chan lp.CCMetric)
if !m.init {
return
}
now := time.Now()
tdiff := now.Sub(m.lastTimestamp)
for device, devData := range m.stats {
stats := m.getDeviceDataCommand(device)
processed := []string{}
for _, line := range stats {
lf := strings.Fields(line)
if len(lf) > 1 {
if fields, ok := m.matches[lf[0]]; ok {
for name, idx := range fields {
x, err := strconv.ParseInt(lf[idx], 0, 64)
if err != nil {
continue
}
value := x - devData[name]
devData[name] = x
if value < 0 {
value = 0
}
y, err := lp.New(name, m.tags, m.meta, map[string]interface{}{"value": value}, time.Now())
if err == nil {
y.AddTag("device", device)
if strings.Contains(name, "byte") {
y.AddMeta("unit", "Byte")
}
output <- y
if m.config.SendAllMetrics {
processed = append(processed, name)
}
}
}
}
data := m.getDeviceDataCommand(device)
for _, def := range m.definitions {
var use_x int64
var err error
var y lp.CCMetric
x, err := getMetricData(data, def.lineprefix, def.lineoffset)
if err == nil {
use_x = x
} else {
use_x = devData[def.name]
}
}
if m.config.SendAllMetrics {
for name := range devData {
if _, done := stringArrayContains(processed, name); !done {
y, err := lp.New(name, m.tags, m.meta, map[string]interface{}{"value": 0}, time.Now())
if err == nil {
y.AddTag("device", device)
if strings.Contains(name, "byte") {
y.AddMeta("unit", "Byte")
}
output <- y
}
var value interface{}
switch def.calc {
case "none":
value = use_x
y, err = lp.New(def.name, m.tags, m.meta, map[string]interface{}{"value": value}, time.Now())
case "difference":
value = use_x - devData[def.name]
if value.(int64) < 0 {
value = 0
}
y, err = lp.New(def.name, m.tags, m.meta, map[string]interface{}{"value": value}, time.Now())
case "derivative":
value = float64(use_x-devData[def.name]) / tdiff.Seconds()
if value.(float64) < 0 {
value = 0
}
y, err = lp.New(def.name, m.tags, m.meta, map[string]interface{}{"value": value}, time.Now())
}
if err == nil {
y.AddTag("device", device)
if len(def.unit) > 0 {
y.AddMeta("unit", def.unit)
}
output <- y
}
devData[def.name] = use_x
}
}
m.lastTimestamp = now
}
func (m *LustreCollector) Close() {

View File

@@ -3,27 +3,44 @@
```json
"lustrestat": {
"procfiles" : [
"/proc/fs/lustre/llite/lnec-XXXXXX/stats"
],
"lctl_command": "/path/to/lctl",
"exclude_metrics": [
"setattr",
"getattr"
]
],
"send_abs_values" : true,
"send_derived_values" : true,
"send_diff_values": true,
"use_sudo": false
}
```
The `lustrestat` collector reads from the procfs stat files for Lustre like `/proc/fs/lustre/llite/lnec-XXXXXX/stats`.
The `lustrestat` collector uses the `lctl` application with the `get_param` option to get all `llite` metrics (Lustre client). The `llite` metrics are only available for root users. If password-less sudo is configured, you can enable `sudo` in the configuration.
Metrics:
* `read_bytes`
* `read_requests`
* `write_bytes`
* `write_requests`
* `open`
* `close`
* `getattr`
* `setattr`
* `statfs`
* `inode_permission`
* `lustre_read_bytes` (unit `bytes`)
* `lustre_read_requests` (unit `requests`)
* `lustre_write_bytes` (unit `bytes`)
* `lustre_write_requests` (unit `requests`)
* `lustre_open`
* `lustre_close`
* `lustre_getattr`
* `lustre_setattr`
* `lustre_statfs`
* `lustre_inode_permission`
* `lustre_read_bw` (if `send_derived_values == true`, unit `bytes/sec`)
* `lustre_write_bw` (if `send_derived_values == true`, unit `bytes/sec`)
* `lustre_read_requests_rate` (if `send_derived_values == true`, unit `requests/sec`)
* `lustre_write_requests_rate` (if `send_derived_values == true`, unit `requests/sec`)
* `lustre_read_bytes_diff` (if `send_diff_values == true`, unit `bytes`)
* `lustre_read_requests_diff` (if `send_diff_values == true`, unit `requests`)
* `lustre_write_bytes_diff` (if `send_diff_values == true`, unit `bytes`)
* `lustre_write_requests_diff` (if `send_diff_values == true`, unit `requests`)
* `lustre_open_diff` (if `send_diff_values == true`)
* `lustre_close_diff` (if `send_diff_values == true`)
* `lustre_getattr_diff` (if `send_diff_values == true`)
* `lustre_setattr_diff` (if `send_diff_values == true`)
* `lustre_statfs_diff` (if `send_diff_values == true`)
* `lustre_inode_permission_diff` (if `send_diff_values == true`)
This collector adds an `device` tag.

View File

@@ -32,15 +32,21 @@ type MemstatCollectorNode struct {
type MemstatCollector struct {
metricCollector
stats map[string]int64
tags map[string]string
matches map[string]string
config MemstatCollectorConfig
nodefiles map[int]MemstatCollectorNode
stats map[string]int64
tags map[string]string
matches map[string]string
config MemstatCollectorConfig
nodefiles map[int]MemstatCollectorNode
sendMemUsed bool
}
func getStats(filename string) map[string]float64 {
stats := make(map[string]float64)
type MemstatStats struct {
value float64
unit string
}
func getStats(filename string) map[string]MemstatStats {
stats := make(map[string]MemstatStats)
file, err := os.Open(filename)
if err != nil {
cclog.Error(err.Error())
@@ -54,12 +60,18 @@ func getStats(filename string) map[string]float64 {
if len(linefields) == 3 {
v, err := strconv.ParseFloat(linefields[1], 64)
if err == nil {
stats[strings.Trim(linefields[0], ":")] = v
stats[strings.Trim(linefields[0], ":")] = MemstatStats{
value: v,
unit: linefields[2],
}
}
} else if len(linefields) == 5 {
v, err := strconv.ParseFloat(linefields[3], 64)
if err == nil {
stats[strings.Trim(linefields[0], ":")] = v
stats[strings.Trim(linefields[0], ":")] = MemstatStats{
value: v,
unit: linefields[4],
}
}
}
}
@@ -69,6 +81,7 @@ func getStats(filename string) map[string]float64 {
func (m *MemstatCollector) Init(config json.RawMessage) error {
var err error
m.name = "MemstatCollector"
m.parallel = true
m.config.NodeStats = true
m.config.NumaStats = false
if len(config) > 0 {
@@ -77,7 +90,7 @@ func (m *MemstatCollector) Init(config json.RawMessage) error {
return err
}
}
m.meta = map[string]string{"source": m.name, "group": "Memory", "unit": "kByte"}
m.meta = map[string]string{"source": m.name, "group": "Memory"}
m.stats = make(map[string]int64)
m.matches = make(map[string]string)
m.tags = map[string]string{"type": "node"}
@@ -99,6 +112,10 @@ func (m *MemstatCollector) Init(config json.RawMessage) error {
m.matches[k] = v
}
}
m.sendMemUsed = false
if _, skip := stringArrayContains(m.config.ExcludeMetrics, "mem_used"); !skip {
m.sendMemUsed = true
}
if len(m.matches) == 0 {
return errors.New("no metrics to collect")
}
@@ -143,32 +160,57 @@ func (m *MemstatCollector) Init(config json.RawMessage) error {
func (m *MemstatCollector) Read(interval time.Duration, output chan lp.CCMetric) {
if !m.init {
cclog.ComponentPrint(m.name, "Here")
return
}
sendStats := func(stats map[string]float64, tags map[string]string) {
sendStats := func(stats map[string]MemstatStats, tags map[string]string) {
for match, name := range m.matches {
var value float64 = 0
var unit string = ""
if v, ok := stats[match]; ok {
value = v
value = v.value
if len(v.unit) > 0 {
unit = v.unit
}
}
y, err := lp.New(name, tags, m.meta, map[string]interface{}{"value": value}, time.Now())
if err == nil {
if len(unit) > 0 {
y.AddMeta("unit", unit)
}
output <- y
}
}
if _, skip := stringArrayContains(m.config.ExcludeMetrics, "mem_used"); !skip {
if freeVal, free := stats["MemFree"]; free {
if bufVal, buffers := stats["Buffers"]; buffers {
if cacheVal, cached := stats["Cached"]; cached {
memUsed := stats["MemTotal"] - (freeVal + bufVal + cacheVal)
y, err := lp.New("mem_used", tags, m.meta, map[string]interface{}{"value": memUsed}, time.Now())
if err == nil {
output <- y
if m.sendMemUsed {
memUsed := 0.0
unit := ""
if totalVal, total := stats["MemTotal"]; total {
if freeVal, free := stats["MemFree"]; free {
if bufVal, buffers := stats["Buffers"]; buffers {
if cacheVal, cached := stats["Cached"]; cached {
memUsed = totalVal.value - (freeVal.value + bufVal.value + cacheVal.value)
if len(totalVal.unit) > 0 {
unit = totalVal.unit
} else if len(freeVal.unit) > 0 {
unit = freeVal.unit
} else if len(bufVal.unit) > 0 {
unit = bufVal.unit
} else if len(cacheVal.unit) > 0 {
unit = cacheVal.unit
}
}
}
}
}
y, err := lp.New("mem_used", tags, m.meta, map[string]interface{}{"value": memUsed}, time.Now())
if err == nil {
if len(unit) > 0 {
y.AddMeta("unit", unit)
}
output <- y
}
}
}

View File

@@ -3,27 +3,25 @@ package collectors
import (
"encoding/json"
"fmt"
"io/ioutil"
"log"
"strconv"
"strings"
"time"
lp "github.com/ClusterCockpit/cc-metric-collector/internal/ccMetric"
)
type MetricCollector interface {
Name() string // Name of the metric collector
Init(config json.RawMessage) error // Initialize metric collector
Initialized() bool // Is metric collector initialized?
Name() string // Name of the metric collector
Init(config json.RawMessage) error // Initialize metric collector
Initialized() bool // Is metric collector initialized?
Parallel() bool
Read(duration time.Duration, output chan lp.CCMetric) // Read metrics from metric collector
Close() // Close / finish metric collector
}
type metricCollector struct {
name string // name of the metric
init bool // is metric collector initialized?
meta map[string]string // static meta data tags
name string // name of the metric
init bool // is metric collector initialized?
parallel bool // can the metric collector be executed in parallel with others
meta map[string]string // static meta data tags
}
// Name returns the name of the metric collector
@@ -31,6 +29,11 @@ func (c *metricCollector) Name() string {
return c.name
}
// Name returns the name of the metric collector
func (c *metricCollector) Parallel() bool {
return c.parallel
}
// Setup is for future use
func (c *metricCollector) setup() error {
return nil
@@ -65,58 +68,6 @@ func stringArrayContains(array []string, str string) (int, bool) {
return -1, false
}
// SocketList returns the list of physical sockets as read from /proc/cpuinfo
func SocketList() []int {
buffer, err := ioutil.ReadFile("/proc/cpuinfo")
if err != nil {
log.Print(err)
return nil
}
ll := strings.Split(string(buffer), "\n")
var packs []int
for _, line := range ll {
if strings.HasPrefix(line, "physical id") {
lv := strings.Fields(line)
id, err := strconv.ParseInt(lv[3], 10, 32)
if err != nil {
log.Print(err)
return packs
}
_, found := intArrayContains(packs, int(id))
if !found {
packs = append(packs, int(id))
}
}
}
return packs
}
// CpuList returns the list of physical CPUs (in contrast to logical CPUs) as read from /proc/cpuinfo
func CpuList() []int {
buffer, err := ioutil.ReadFile("/proc/cpuinfo")
if err != nil {
log.Print(err)
return nil
}
ll := strings.Split(string(buffer), "\n")
var cpulist []int
for _, line := range ll {
if strings.HasPrefix(line, "processor") {
lv := strings.Fields(line)
id, err := strconv.ParseInt(lv[2], 10, 32)
if err != nil {
log.Print(err)
return cpulist
}
_, found := intArrayContains(cpulist, int(id))
if !found {
cpulist = append(cpulist, int(id))
}
}
}
return cpulist
}
// RemoveFromStringList removes the string r from the array of strings s
// If r is not contained in the array an error is returned
func RemoveFromStringList(s []string, r string) ([]string, error) {
@@ -125,5 +76,5 @@ func RemoveFromStringList(s []string, r string) ([]string, error) {
return append(s[:i], s[i+1:]...), nil
}
}
return s, fmt.Errorf("No such string in list")
return s, fmt.Errorf("no such string in list")
}

View File

@@ -13,38 +13,62 @@ import (
lp "github.com/ClusterCockpit/cc-metric-collector/internal/ccMetric"
)
const NETSTATFILE = `/proc/net/dev`
const NETSTATFILE = "/proc/net/dev"
type NetstatCollectorConfig struct {
IncludeDevices []string `json:"include_devices"`
IncludeDevices []string `json:"include_devices"`
SendAbsoluteValues bool `json:"send_abs_values"`
SendDerivedValues bool `json:"send_derived_values"`
}
type NetstatCollectorMetric struct {
index int
lastValue float64
name string
index int
tags map[string]string
meta map[string]string
meta_rates map[string]string
lastValue int64
}
type NetstatCollector struct {
metricCollector
config NetstatCollectorConfig
matches map[string]map[string]NetstatCollectorMetric
devtags map[string]map[string]string
matches map[string][]NetstatCollectorMetric
lastTimestamp time.Time
}
func (m *NetstatCollector) Init(config json.RawMessage) error {
m.name = "NetstatCollector"
m.parallel = true
m.setup()
m.lastTimestamp = time.Now()
m.meta = map[string]string{"source": m.name, "group": "Network"}
m.devtags = make(map[string]map[string]string)
nameIndexMap := map[string]int{
"net_bytes_in": 1,
"net_pkts_in": 2,
"net_bytes_out": 9,
"net_pkts_out": 10,
}
m.matches = make(map[string]map[string]NetstatCollectorMetric)
const (
fieldInterface = iota
fieldReceiveBytes
fieldReceivePackets
fieldReceiveErrs
fieldReceiveDrop
fieldReceiveFifo
fieldReceiveFrame
fieldReceiveCompressed
fieldReceiveMulticast
fieldTransmitBytes
fieldTransmitPackets
fieldTransmitErrs
fieldTransmitDrop
fieldTransmitFifo
fieldTransmitColls
fieldTransmitCarrier
fieldTransmitCompressed
)
m.matches = make(map[string][]NetstatCollectorMetric)
// Set default configuration,
m.config.SendAbsoluteValues = true
m.config.SendDerivedValues = false
// Read configuration file, allow overwriting default config
if len(config) > 0 {
err := json.Unmarshal(config, &m.config)
if err != nil {
@@ -52,7 +76,9 @@ func (m *NetstatCollector) Init(config json.RawMessage) error {
return err
}
}
file, err := os.Open(string(NETSTATFILE))
// Check access to net statistic file
file, err := os.Open(NETSTATFILE)
if err != nil {
cclog.ComponentError(m.name, err.Error())
return err
@@ -62,23 +88,65 @@ func (m *NetstatCollector) Init(config json.RawMessage) error {
scanner := bufio.NewScanner(file)
for scanner.Scan() {
l := scanner.Text()
// Skip lines with no net device entry
if !strings.Contains(l, ":") {
continue
}
// Split line into fields
f := strings.Fields(l)
// Get net device entry
dev := strings.Trim(f[0], ": ")
// Check if device is a included device
if _, ok := stringArrayContains(m.config.IncludeDevices, dev); ok {
m.matches[dev] = make(map[string]NetstatCollectorMetric)
for name, idx := range nameIndexMap {
m.matches[dev][name] = NetstatCollectorMetric{
index: idx,
lastValue: 0,
}
tags := map[string]string{"device": dev, "type": "node"}
meta_unit_byte := map[string]string{"source": m.name, "group": "Network", "unit": "bytes"}
meta_unit_byte_per_sec := map[string]string{"source": m.name, "group": "Network", "unit": "bytes/sec"}
meta_unit_pkts := map[string]string{"source": m.name, "group": "Network", "unit": "packets"}
meta_unit_pkts_per_sec := map[string]string{"source": m.name, "group": "Network", "unit": "packets/sec"}
m.matches[dev] = []NetstatCollectorMetric{
{
name: "net_bytes_in",
index: fieldReceiveBytes,
lastValue: -1,
tags: tags,
meta: meta_unit_byte,
meta_rates: meta_unit_byte_per_sec,
},
{
name: "net_pkts_in",
index: fieldReceivePackets,
lastValue: -1,
tags: tags,
meta: meta_unit_pkts,
meta_rates: meta_unit_pkts_per_sec,
},
{
name: "net_bytes_out",
index: fieldTransmitBytes,
lastValue: -1,
tags: tags,
meta: meta_unit_byte,
meta_rates: meta_unit_byte_per_sec,
},
{
name: "net_pkts_out",
index: fieldTransmitPackets,
lastValue: -1,
tags: tags,
meta: meta_unit_pkts,
meta_rates: meta_unit_pkts_per_sec,
},
}
m.devtags[dev] = map[string]string{"device": dev, "type": "node"}
}
}
if len(m.devtags) == 0 {
if len(m.matches) == 0 {
return errors.New("no devices to collector metrics found")
}
m.init = true
@@ -89,50 +157,62 @@ func (m *NetstatCollector) Read(interval time.Duration, output chan lp.CCMetric)
if !m.init {
return
}
// Current time stamp
now := time.Now()
// time difference to last time stamp
timeDiff := now.Sub(m.lastTimestamp).Seconds()
// Save current timestamp
m.lastTimestamp = now
file, err := os.Open(string(NETSTATFILE))
if err != nil {
cclog.ComponentError(m.name, err.Error())
return
}
defer file.Close()
tdiff := now.Sub(m.lastTimestamp)
scanner := bufio.NewScanner(file)
for scanner.Scan() {
l := scanner.Text()
// Skip lines with no net device entry
if !strings.Contains(l, ":") {
continue
}
// Split line into fields
f := strings.Fields(l)
// Get net device entry
dev := strings.Trim(f[0], ":")
// Check if device is a included device
if devmetrics, ok := m.matches[dev]; ok {
for name, data := range devmetrics {
v, err := strconv.ParseFloat(f[data.index], 64)
if err == nil {
vdiff := v - data.lastValue
value := vdiff / tdiff.Seconds()
if data.lastValue == 0 {
value = 0
}
data.lastValue = v
y, err := lp.New(name, m.devtags[dev], m.meta, map[string]interface{}{"value": value}, now)
if err == nil {
switch {
case strings.Contains(name, "byte"):
y.AddMeta("unit", "bytes/sec")
case strings.Contains(name, "pkt"):
y.AddMeta("unit", "packets/sec")
}
for i := range devmetrics {
metric := &devmetrics[i]
// Read value
v, err := strconv.ParseInt(f[metric.index], 10, 64)
if err != nil {
continue
}
if m.config.SendAbsoluteValues {
if y, err := lp.New(metric.name, metric.tags, metric.meta, map[string]interface{}{"value": v}, now); err == nil {
output <- y
}
devmetrics[name] = data
}
if m.config.SendDerivedValues {
if metric.lastValue >= 0 {
rate := float64(v-metric.lastValue) / timeDiff
if y, err := lp.New(metric.name+"_bw", metric.tags, metric.meta_rates, map[string]interface{}{"value": rate}, now); err == nil {
output <- y
}
}
metric.lastValue = v
}
}
}
}
m.lastTimestamp = time.Now()
}
func (m *NetstatCollector) Close() {

View File

@@ -5,17 +5,23 @@
"netstat": {
"include_devices": [
"eth0"
]
],
"send_abs_values" : true,
"send_derived_values" : true
}
```
The `netstat` collector reads data from `/proc/net/dev` and outputs a handful **node** metrics. With the `include_devices` list you can specify which network devices should be measured. **Note**: Most other collectors use an _exclude_ list instead of an include list.
Metrics:
* `net_bytes_in` (`unit=bytes/sec`)
* `net_bytes_out` (`unit=bytes/sec`)
* `net_pkts_in` (`unit=packets/sec`)
* `net_pkts_out` (`unit=packets/sec`)
* `net_bytes_in` (`unit=bytes`)
* `net_bytes_out` (`unit=bytes`)
* `net_pkts_in` (`unit=packets`)
* `net_pkts_out` (`unit=packets`)
* `net_bytes_in_bw` (`unit=bytes/sec` if `send_derived_values == true`)
* `net_bytes_out_bw` (`unit=bytes/sec` if `send_derived_values == true`)
* `net_pkts_in_bw` (`unit=packets/sec` if `send_derived_values == true`)
* `net_pkts_out_bw` (`unit=packets/sec` if `send_derived_values == true`)
The device name is added as tag `device`.

View File

@@ -36,7 +36,7 @@ type nfsCollector struct {
}
func (m *nfsCollector) initStats() error {
cmd := exec.Command(m.config.Nfsstats, `-l`)
cmd := exec.Command(m.config.Nfsstats, `-l`, `--all`)
cmd.Wait()
buffer, err := cmd.Output()
if err == nil {
@@ -52,7 +52,7 @@ func (m *nfsCollector) initStats() error {
if err == nil {
x := m.data[name]
x.current = value
x.last = 0
x.last = value
m.data[name] = x
}
}
@@ -63,7 +63,7 @@ func (m *nfsCollector) initStats() error {
}
func (m *nfsCollector) updateStats() error {
cmd := exec.Command(m.config.Nfsstats, `-l`)
cmd := exec.Command(m.config.Nfsstats, `-l`, `--all`)
cmd.Wait()
buffer, err := cmd.Output()
if err == nil {
@@ -114,6 +114,7 @@ func (m *nfsCollector) MainInit(config json.RawMessage) error {
m.data = make(map[string]NfsCollectorData)
m.initStats()
m.init = true
m.parallel = true
return nil
}

View File

@@ -54,6 +54,7 @@ func (m *NUMAStatsCollector) Init(config json.RawMessage) error {
}
m.name = "NUMAStatsCollector"
m.parallel = true
m.setup()
m.meta = map[string]string{
"source": m.name,

File diff suppressed because it is too large Load Diff

View File

@@ -3,38 +3,74 @@
```json
"nvidia": {
"exclude_devices" : [
"0","1"
"exclude_devices": [
"0","1", "0000000:ff:01.0"
],
"exclude_metrics": [
"nv_fb_memory",
"nv_fb_mem_used",
"nv_fan"
]
],
"process_mig_devices": false,
"use_pci_info_as_type_id": true,
"add_pci_info_tag": false,
"add_uuid_meta": false,
"add_board_number_meta": false,
"add_serial_meta": false,
"use_uuid_for_mig_device": false,
"use_slice_for_mig_device": false
}
```
The `nvidia` collector can be configured to leave out specific devices with the `exclude_devices` option. It takes IDs as supplied to the NVML with `nvmlDeviceGetHandleByIndex()` or the PCI address in NVML format (`%08X:%02X:%02X.0`). Metrics (listed below) that should not be sent to the MetricRouter can be excluded with the `exclude_metrics` option. Commonly only the physical GPUs are monitored. If MIG devices should be analyzed as well, set `process_mig_devices` (adds `stype=mig,stype-id=<mig_index>`). With the options `use_uuid_for_mig_device` and `use_slice_for_mig_device`, the `<mig_index>` can be replaced with the UUID (e.g. `MIG-6a9f7cc8-6d5b-5ce0-92de-750edc4d8849`) or the MIG slice name (e.g. `1g.5gb`).
The metrics sent by the `nvidia` collector use `accelerator` as `type` tag. For the `type-id`, it uses the device handle index by default. With the `use_pci_info_as_type_id` option, the PCI ID is used instead. If both values should be added as tags, activate the `add_pci_info_tag` option. It uses the device handle index as `type-id` and adds the PCI ID as separate `pci_identifier` tag.
Optionally, it is possible to add the UUID, the board part number and the serial to the meta informations. They are not sent to the sinks (if not configured otherwise).
Metrics:
* `nv_util`
* `nv_mem_util`
* `nv_mem_total`
* `nv_fb_memory`
* `nv_fb_mem_total`
* `nv_fb_mem_used`
* `nv_bar1_mem_total`
* `nv_bar1_mem_used`
* `nv_temp`
* `nv_fan`
* `nv_ecc_mode`
* `nv_perf_state`
* `nv_power_usage_report`
* `nv_graphics_clock_report`
* `nv_sm_clock_report`
* `nv_mem_clock_report`
* `nv_power_usage`
* `nv_graphics_clock`
* `nv_sm_clock`
* `nv_mem_clock`
* `nv_video_clock`
* `nv_max_graphics_clock`
* `nv_max_sm_clock`
* `nv_max_mem_clock`
* `nv_ecc_db_error`
* `nv_ecc_sb_error`
* `nv_power_man_limit`
* `nv_max_video_clock`
* `nv_ecc_uncorrected_error`
* `nv_ecc_corrected_error`
* `nv_power_max_limit`
* `nv_encoder_util`
* `nv_decoder_util`
* `nv_remapped_rows_corrected`
* `nv_remapped_rows_uncorrected`
* `nv_remapped_rows_pending`
* `nv_remapped_rows_failure`
* `nv_compute_processes`
* `nv_graphics_processes`
* `nv_violation_power`
* `nv_violation_thermal`
* `nv_violation_sync_boost`
* `nv_violation_board_limit`
* `nv_violation_low_util`
* `nv_violation_reliability`
* `nv_violation_below_app_clock`
* `nv_violation_below_base_clock`
* `nv_nvlink_crc_flit_errors`
* `nv_nvlink_crc_errors`
* `nv_nvlink_ecc_errors`
* `nv_nvlink_replay_errors`
* `nv_nvlink_recovery_errors`
It uses a separate `type` in the metrics. The output metric looks like this:
`<name>,type=accelerator,type-id=<nvidia-gpu-id> value=<metric value> <timestamp>`
Some metrics add the additional sub type tag (`stype`) like the `nv_nvlink_*` metrics set `stype=nvlink,stype-id=<link_number>`.

319
collectors/rocmsmiMetric.go Normal file
View File

@@ -0,0 +1,319 @@
package collectors
import (
"encoding/json"
"errors"
"fmt"
"time"
cclog "github.com/ClusterCockpit/cc-metric-collector/internal/ccLogger"
lp "github.com/ClusterCockpit/cc-metric-collector/internal/ccMetric"
"github.com/ClusterCockpit/go-rocm-smi/pkg/rocm_smi"
)
type RocmSmiCollectorConfig struct {
ExcludeMetrics []string `json:"exclude_metrics,omitempty"`
ExcludeDevices []string `json:"exclude_devices,omitempty"`
AddPciInfoTag bool `json:"add_pci_info_tag,omitempty"`
UsePciInfoAsTypeId bool `json:"use_pci_info_as_type_id,omitempty"`
AddSerialMeta bool `json:"add_serial_meta,omitempty"`
}
type RocmSmiCollectorDevice struct {
device rocm_smi.DeviceHandle
index int
tags map[string]string // default tags
meta map[string]string // default meta information
excludeMetrics map[string]bool // copy of exclude metrics from config
}
type RocmSmiCollector struct {
metricCollector
config RocmSmiCollectorConfig // the configuration structure
devices []RocmSmiCollectorDevice
}
// Functions to implement MetricCollector interface
// Init(...), Read(...), Close()
// See: metricCollector.go
// Init initializes the sample collector
// Called once by the collector manager
// All tags, meta data tags and metrics that do not change over the runtime should be set here
func (m *RocmSmiCollector) Init(config json.RawMessage) error {
var err error = nil
// Always set the name early in Init() to use it in cclog.Component* functions
m.name = "RocmSmiCollector"
// This is for later use, also call it early
m.setup()
// Define meta information sent with each metric
// (Can also be dynamic or this is the basic set with extension through AddMeta())
//m.meta = map[string]string{"source": m.name, "group": "AMD"}
// Define tags sent with each metric
// The 'type' tag is always needed, it defines the granulatity of the metric
// node -> whole system
// socket -> CPU socket (requires socket ID as 'type-id' tag)
// cpu -> single CPU hardware thread (requires cpu ID as 'type-id' tag)
//m.tags = map[string]string{"type": "node"}
// Read in the JSON configuration
if len(config) > 0 {
err = json.Unmarshal(config, &m.config)
if err != nil {
cclog.ComponentError(m.name, "Error reading config:", err.Error())
return err
}
}
ret := rocm_smi.Init()
if ret != rocm_smi.STATUS_SUCCESS {
err = errors.New("Failed to initialize ROCm SMI library")
cclog.ComponentError(m.name, err.Error())
return err
}
numDevs, ret := rocm_smi.NumMonitorDevices()
if ret != rocm_smi.STATUS_SUCCESS {
err = errors.New("Failed to get number of GPUs from ROCm SMI library")
cclog.ComponentError(m.name, err.Error())
return err
}
exclDev := func(s string) bool {
skip_device := false
for _, excl := range m.config.ExcludeDevices {
if excl == s {
skip_device = true
break
}
}
return skip_device
}
m.devices = make([]RocmSmiCollectorDevice, 0)
for i := 0; i < numDevs; i++ {
str_i := fmt.Sprintf("%d", i)
if exclDev(str_i) {
continue
}
device, ret := rocm_smi.DeviceGetHandleByIndex(i)
if ret != rocm_smi.STATUS_SUCCESS {
err = fmt.Errorf("Failed to get handle for GPU %d", i)
cclog.ComponentError(m.name, err.Error())
return err
}
pciInfo, ret := rocm_smi.DeviceGetPciInfo(device)
if ret != rocm_smi.STATUS_SUCCESS {
err = fmt.Errorf("Failed to get PCI information for GPU %d", i)
cclog.ComponentError(m.name, err.Error())
return err
}
pciId := fmt.Sprintf(
"%08X:%02X:%02X.%X",
pciInfo.Domain,
pciInfo.Bus,
pciInfo.Device,
pciInfo.Function)
if exclDev(pciId) {
continue
}
dev := RocmSmiCollectorDevice{
device: device,
tags: map[string]string{
"type": "accelerator",
"type-id": str_i,
},
meta: map[string]string{
"source": m.name,
"group": "AMD",
},
}
if m.config.UsePciInfoAsTypeId {
dev.tags["type-id"] = pciId
} else if m.config.AddPciInfoTag {
dev.tags["pci_identifier"] = pciId
}
if m.config.AddSerialMeta {
serial, ret := rocm_smi.DeviceGetSerialNumber(device)
if ret != rocm_smi.STATUS_SUCCESS {
cclog.ComponentError(m.name, "Unable to get serial number for device at index", i, ":", rocm_smi.StatusStringNoError(ret))
} else {
dev.meta["serial"] = serial
}
}
// Add excluded metrics
dev.excludeMetrics = map[string]bool{}
for _, e := range m.config.ExcludeMetrics {
dev.excludeMetrics[e] = true
}
dev.index = i
m.devices = append(m.devices, dev)
}
// Set this flag only if everything is initialized properly, all required files exist, ...
m.init = true
return err
}
// Read collects all metrics belonging to the sample collector
// and sends them through the output channel to the collector manager
func (m *RocmSmiCollector) Read(interval time.Duration, output chan lp.CCMetric) {
// Create a sample metric
timestamp := time.Now()
for _, dev := range m.devices {
metrics, ret := rocm_smi.DeviceGetMetrics(dev.device)
if ret != rocm_smi.STATUS_SUCCESS {
cclog.ComponentError(m.name, "Unable to get metrics for device at index", dev.index, ":", rocm_smi.StatusStringNoError(ret))
continue
}
if !dev.excludeMetrics["rocm_gfx_util"] {
value := metrics.Average_gfx_activity
y, err := lp.New("rocm_gfx_util", dev.tags, dev.meta, map[string]interface{}{"value": value}, timestamp)
if err == nil {
output <- y
}
}
if !dev.excludeMetrics["rocm_umc_util"] {
value := metrics.Average_umc_activity
y, err := lp.New("rocm_umc_util", dev.tags, dev.meta, map[string]interface{}{"value": value}, timestamp)
if err == nil {
output <- y
}
}
if !dev.excludeMetrics["rocm_mm_util"] {
value := metrics.Average_mm_activity
y, err := lp.New("rocm_mm_util", dev.tags, dev.meta, map[string]interface{}{"value": value}, timestamp)
if err == nil {
output <- y
}
}
if !dev.excludeMetrics["rocm_avg_power"] {
value := metrics.Average_socket_power
y, err := lp.New("rocm_avg_power", dev.tags, dev.meta, map[string]interface{}{"value": value}, timestamp)
if err == nil {
output <- y
}
}
if !dev.excludeMetrics["rocm_temp_mem"] {
value := metrics.Temperature_mem
y, err := lp.New("rocm_temp_mem", dev.tags, dev.meta, map[string]interface{}{"value": value}, timestamp)
if err == nil {
output <- y
}
}
if !dev.excludeMetrics["rocm_temp_hotspot"] {
value := metrics.Temperature_hotspot
y, err := lp.New("rocm_temp_hotspot", dev.tags, dev.meta, map[string]interface{}{"value": value}, timestamp)
if err == nil {
output <- y
}
}
if !dev.excludeMetrics["rocm_temp_edge"] {
value := metrics.Temperature_edge
y, err := lp.New("rocm_temp_edge", dev.tags, dev.meta, map[string]interface{}{"value": value}, timestamp)
if err == nil {
output <- y
}
}
if !dev.excludeMetrics["rocm_temp_vrgfx"] {
value := metrics.Temperature_vrgfx
y, err := lp.New("rocm_temp_vrgfx", dev.tags, dev.meta, map[string]interface{}{"value": value}, timestamp)
if err == nil {
output <- y
}
}
if !dev.excludeMetrics["rocm_temp_vrsoc"] {
value := metrics.Temperature_vrsoc
y, err := lp.New("rocm_temp_vrsoc", dev.tags, dev.meta, map[string]interface{}{"value": value}, timestamp)
if err == nil {
output <- y
}
}
if !dev.excludeMetrics["rocm_temp_vrmem"] {
value := metrics.Temperature_vrmem
y, err := lp.New("rocm_temp_vrmem", dev.tags, dev.meta, map[string]interface{}{"value": value}, timestamp)
if err == nil {
output <- y
}
}
if !dev.excludeMetrics["rocm_gfx_clock"] {
value := metrics.Average_gfxclk_frequency
y, err := lp.New("rocm_gfx_clock", dev.tags, dev.meta, map[string]interface{}{"value": value}, timestamp)
if err == nil {
output <- y
}
}
if !dev.excludeMetrics["rocm_soc_clock"] {
value := metrics.Average_socclk_frequency
y, err := lp.New("rocm_soc_clock", dev.tags, dev.meta, map[string]interface{}{"value": value}, timestamp)
if err == nil {
output <- y
}
}
if !dev.excludeMetrics["rocm_u_clock"] {
value := metrics.Average_uclk_frequency
y, err := lp.New("rocm_u_clock", dev.tags, dev.meta, map[string]interface{}{"value": value}, timestamp)
if err == nil {
output <- y
}
}
if !dev.excludeMetrics["rocm_v0_clock"] {
value := metrics.Average_vclk0_frequency
y, err := lp.New("rocm_v0_clock", dev.tags, dev.meta, map[string]interface{}{"value": value}, timestamp)
if err == nil {
output <- y
}
}
if !dev.excludeMetrics["rocm_v1_clock"] {
value := metrics.Average_vclk1_frequency
y, err := lp.New("rocm_v1_clock", dev.tags, dev.meta, map[string]interface{}{"value": value}, timestamp)
if err == nil {
output <- y
}
}
if !dev.excludeMetrics["rocm_d0_clock"] {
value := metrics.Average_dclk0_frequency
y, err := lp.New("rocm_d0_clock", dev.tags, dev.meta, map[string]interface{}{"value": value}, timestamp)
if err == nil {
output <- y
}
}
if !dev.excludeMetrics["rocm_d1_clock"] {
value := metrics.Average_dclk1_frequency
y, err := lp.New("rocm_d1_clock", dev.tags, dev.meta, map[string]interface{}{"value": value}, timestamp)
if err == nil {
output <- y
}
}
if !dev.excludeMetrics["rocm_temp_hbm"] {
for i := 0; i < rocm_smi.NUM_HBM_INSTANCES; i++ {
value := metrics.Temperature_hbm[i]
y, err := lp.New("rocm_temp_hbm", dev.tags, dev.meta, map[string]interface{}{"value": value}, timestamp)
if err == nil {
y.AddTag("stype", "device")
y.AddTag("stype-id", fmt.Sprintf("%d", i))
output <- y
}
}
}
}
}
// Close metric collector: close network connection, close files, close libraries, ...
// Called once by the collector manager
func (m *RocmSmiCollector) Close() {
// Unset flag
ret := rocm_smi.Shutdown()
if ret != rocm_smi.STATUS_SUCCESS {
cclog.ComponentError(m.name, "Failed to shutdown ROCm SMI library")
}
m.init = false
}

View File

@@ -0,0 +1,47 @@
## `rocm_smi` collector
```json
"rocm_smi": {
"exclude_devices": [
"0","1", "0000000:ff:01.0"
],
"exclude_metrics": [
"rocm_mm_util",
"rocm_temp_vrsoc"
],
"use_pci_info_as_type_id": true,
"add_pci_info_tag": false,
"add_serial_meta": false,
}
```
The `rocm_smi` collector can be configured to leave out specific devices with the `exclude_devices` option. It takes logical IDs in the list of available devices or the PCI address similar to NVML format (`%08X:%02X:%02X.0`). Metrics (listed below) that should not be sent to the MetricRouter can be excluded with the `exclude_metrics` option.
The metrics sent by the `rocm_smi` collector use `accelerator` as `type` tag. For the `type-id`, it uses the device handle index by default. With the `use_pci_info_as_type_id` option, the PCI ID is used instead. If both values should be added as tags, activate the `add_pci_info_tag` option. It uses the device handle index as `type-id` and adds the PCI ID as separate `pci_identifier` tag.
Optionally, it is possible to add the serial to the meta informations. They are not sent to the sinks (if not configured otherwise).
Metrics:
* `rocm_gfx_util`
* `rocm_umc_util`
* `rocm_mm_util`
* `rocm_avg_power`
* `rocm_temp_mem`
* `rocm_temp_hotspot`
* `rocm_temp_edge`
* `rocm_temp_vrgfx`
* `rocm_temp_vrsoc`
* `rocm_temp_vrmem`
* `rocm_gfx_clock`
* `rocm_soc_clock`
* `rocm_u_clock`
* `rocm_v0_clock`
* `rocm_v1_clock`
* `rocm_d0_clock`
* `rocm_d1_clock`
* `rocm_temp_hbm`
Some metrics add the additional sub type tag (`stype`) like the `rocm_temp_hbm` metrics set `stype=device,stype-id=<HBM_slice_number>`.

View File

@@ -35,6 +35,10 @@ func (m *SampleCollector) Init(config json.RawMessage) error {
m.name = "InternalCollector"
// This is for later use, also call it early
m.setup()
// Tell whether the collector should be run in parallel with others (reading files, ...)
// or it should be run serially, mostly for collectors acutally doing measurements
// because they should not measure the execution of the other collectors
m.parallel = true
// Define meta information sent with each metric
// (Can also be dynamic or this is the basic set with extension through AddMeta())
m.meta = map[string]string{"source": m.name, "group": "SAMPLE"}
@@ -42,7 +46,12 @@ func (m *SampleCollector) Init(config json.RawMessage) error {
// The 'type' tag is always needed, it defines the granulatity of the metric
// node -> whole system
// socket -> CPU socket (requires socket ID as 'type-id' tag)
// cpu -> single CPU hardware thread (requires cpu ID as 'type-id' tag)
// die -> CPU die (requires CPU die ID as 'type-id' tag)
// memoryDomain -> NUMA domain (requires NUMA domain ID as 'type-id' tag)
// llc -> Last level cache (requires last level cache ID as 'type-id' tag)
// core -> single CPU core that may consist of multiple hardware threads (SMT) (requires core ID as 'type-id' tag)
// hwthtread -> single CPU hardware thread (requires hardware thread ID as 'type-id' tag)
// accelerator -> A accelerator device like GPU or FPGA (requires an accelerator ID as 'type-id' tag)
m.tags = map[string]string{"type": "node"}
// Read in the JSON configuration
if len(config) > 0 {

View File

@@ -50,6 +50,7 @@ func (m *TempCollector) Init(config json.RawMessage) error {
}
m.name = "TempCollector"
m.parallel = true
m.setup()
if len(config) > 0 {
err := json.Unmarshal(config, &m.config)
@@ -70,10 +71,10 @@ func (m *TempCollector) Init(config json.RawMessage) error {
globPattern := filepath.Join("/sys/class/hwmon", "*", "temp*_input")
inputFiles, err := filepath.Glob(globPattern)
if err != nil {
return fmt.Errorf("Unable to glob files with pattern '%s': %v", globPattern, err)
return fmt.Errorf("unable to glob files with pattern '%s': %v", globPattern, err)
}
if inputFiles == nil {
return fmt.Errorf("Unable to find any files with pattern '%s'", globPattern)
return fmt.Errorf("unable to find any files with pattern '%s'", globPattern)
}
// Get sensor name for each temperature sensor file
@@ -116,6 +117,10 @@ func (m *TempCollector) Init(config json.RawMessage) error {
}
// Sensor file
_, err = ioutil.ReadFile(file)
if err != nil {
continue
}
sensor.file = file
// Sensor tags
@@ -158,7 +163,7 @@ func (m *TempCollector) Init(config json.RawMessage) error {
// Empty sensors map
if len(m.sensors) == 0 {
return fmt.Errorf("No temperature sensors found")
return fmt.Errorf("no temperature sensors found")
}
// Finished initialization

View File

@@ -28,6 +28,7 @@ type TopProcsCollector struct {
func (m *TopProcsCollector) Init(config json.RawMessage) error {
var err error
m.name = "TopProcsCollector"
m.parallel = true
m.tags = map[string]string{"type": "node"}
m.meta = map[string]string{"source": m.name, "group": "TopProcs"}
if len(config) > 0 {
@@ -39,14 +40,14 @@ func (m *TopProcsCollector) Init(config json.RawMessage) error {
m.config.Num_procs = int(DEFAULT_NUM_PROCS)
}
if m.config.Num_procs <= 0 || m.config.Num_procs > MAX_NUM_PROCS {
return errors.New(fmt.Sprintf("num_procs option must be set in 'topprocs' config (range: 1-%d)", MAX_NUM_PROCS))
return fmt.Errorf("num_procs option must be set in 'topprocs' config (range: 1-%d)", MAX_NUM_PROCS)
}
m.setup()
command := exec.Command("ps", "-Ao", "comm", "--sort=-pcpu")
command.Wait()
_, err = command.Output()
if err != nil {
return errors.New("Failed to execute command")
return errors.New("failed to execute command")
}
m.init = true
return nil

View File

@@ -1,8 +1,8 @@
{
"sinks": "sinks.json",
"collectors" : "collectors.json",
"receivers" : "receivers.json",
"router" : "router.json",
"interval": 10,
"duration": 1
"sinks": "./sinks.json",
"collectors" : "./collectors.json",
"receivers" : "./receivers.json",
"router" : "./router.json",
"interval": "10s",
"duration": "1s"
}

189
docs/configuration.md Normal file
View File

@@ -0,0 +1,189 @@
# Configuring the CC metric collector
The configuration of the CC metric collector consists of five configuration files: one global file and four component related files.
## Global configuration
The global file contains the paths to the other four files and some global options.
```json
{
"sinks": "sinks.json",
"collectors" : "collectors.json",
"receivers" : "receivers.json",
"router" : "router.json",
"interval": 10,
"duration": 1
}
```
Be aware that the paths are relative to the execution folder of the cc-metric-collector binary, so it is recommended to use absolute paths.
## Component configuration
The others are mainly list of of subcomponents: the collectors, the receivers, the router and the sinks. Their role is best shown in a picture:
```mermaid
flowchart LR
subgraph col ["Collectors"]
direction TB
cpustat["cpustat"]
memstat["memstat"]
tempstat["tempstat"]
misc["..."]
end
subgraph Receivers ["Receivers"]
direction TB
nats["NATS"]
httprecv["HTTP"]
miscrecv[...]
end
subgraph calc["Aggregator"]
direction LR
cache["Cache"]
agg["Calculator"]
end
subgraph sinks ["Sinks"]
direction RL
influx["InfluxDB"]
ganglia["Ganglia"]
logger["Logfile"]
miscsink["..."]
end
cpustat --> CollectorManager["CollectorManager"]
memstat --> CollectorManager
tempstat --> CollectorManager
misc --> CollectorManager
nats --> ReceiverManager["ReceiverManager"]
httprecv --> ReceiverManager
miscrecv --> ReceiverManager
CollectorManager --> newrouter["Router"]
ReceiverManager -.-> newrouter
calc -.-> newrouter
newrouter --> SinkManager["SinkManager"]
newrouter -.-> calc
SinkManager --> influx
SinkManager --> ganglia
SinkManager --> logger
SinkManager --> miscsink
```
There are four parts:
- The collectors read data from files, execute commands and call dynamically loaded library function and send it to the router
- The router can process metrics by cacheing and evaluating functions and conditions on them
- The sinks send the metrics to storage backends
- The receivers can be used to receive metrics from other collectors and forward them to the router. They can be used to create a tree-like structure of collectors.
(A maybe better differentiation between collectors and receivers is that the collectors are called periodically while the receivers have their own logic and submit metrics at any time)
### Collectors configuration file
The collectors configuration file tells which metrics should be queried from the system. The metric gathering is logically grouped in so called 'Collectors'. So there are Collectors to read CPU, memory or filesystem statistics. The collectors configuration file is a list of these collectors with collector-specific configurations:
```json
{
"cpustat" : {},
"diskstat": {
"exclude_metrics": [
"disk_total"
]
}
}
```
The first one is the CPU statistics collector without any collector-specific setting. The second one enables disk mount statistics but excludes the metric `disk_total`.
All names and possible collector-specific configuration options can be found [here](../collectors/README.md).
Some collectors might dynamically load shared libraries. In order to enable these collectors, make sure that the shared library path is part of the `LD_LIBRARY_PATH` environment variable.
### Sinks configuration file
The sinks define the output/sending of metrics. The metrics can be forwarded to multiple sinks, even to sinks of the same type. The sinks configuration file is a list of these sinks, each with an individual name.
```json
{
"myinflux" : {
"type" : "influxasync",
"host": "localhost",
"port": "8086",
"organization" : "testorga",
"database" : "testbucket",
"password" : "<my secret JWT>"
},
"companyinflux" : {
"type" : "influxasync",
"host": "companyhost",
"port": "8086",
"organization" : "company",
"database" : "main",
"password" : "<company's secret JWT>"
}
}
```
The above example configuration file defines two sink, both ot type `influxasync`. They are differentiated internally by the names: `myinflux` and `companyinflux`.
All types and possible sink-specific configuration options can be found [here](../sinks/README.md).
Some sinks might dynamically load shared libraries. In order to enable these sinks, make sure that the shared library path is part of the `LD_LIBRARY_PATH` environment variable.
### Router configuration file
The collectors and the sinks are connected through the router. The router forwards the metrics to the sinks but enables some data processing. A common example is to tag all passing metrics like adding `cluster=mycluster`. But also aggregations like "take the average of all 'ipc' metrics" (ipc -> Instructions Per Cycle). Since the configurations of these aggregations can be quite complicated, we refer to the router's [README](../internal/metricRouter/README.md).
A simple router configuration file to start with looks like this:
```json
{
"add_tags" : [
{
"key" : "cluster",
"value" : "mycluster",
"if" : "*"
}
],
"interval_timestamp" : false,
"num_cache_intervals" : 0
}
```
With the `add_tags` section, we tell to attach the `cluster=mycluster` tag to each (`*` metric). The `interval_timestamp` tell the router to not touch the timestamp of metrics. It is possible to send all metrics within an interval with a common time stamp to avoid later alignment issues. The `num_cache_intervals` diables the cache completely. The cache is only required if you want to do complex metric aggregations.
All configuration options can be found [here](../internal/metricRouter/README.md).
### Receivers configuration file
The receivers are a special feature of the CC Metric Collector to enable simpler integration into exising setups. While collectors query data from the local system, the receivers commonly get data from other systems through some network technology like HTTP or NATS. The idea is keep the current setup but send it to a CC Metric Collector which forwards it to the the destination system (if a sink exists for it). For most setups, the receivers are not required and an the receiver config file should contain only an empty JSON map (`{}`).
```json
{
"nats_rack0": {
"type": "nats",
"address" : "nats-server.example.org",
"port" : "4222",
"subject" : "rack0",
},
"nats_rack1": {
"type": "nats",
"address" : "nats-server.example.org",
"port" : "4222",
"subject" : "rack1",
}
}
```
This example configuration creates two receivers with the names `nats_rack0` and `nats_rack1`. While one subscribes to metrics published with the `rack0` subject, the other one subscribes to the `rack0` subject. The NATS server is the same as it manages all subjects in a subnet. (As example, the router could add tags `rack=0` and `rack=1` respectively to the received metrics.)
All types and possible receiver-specific configuration options can be found [here](../receivers/README.md).

42
go.mod
View File

@@ -1,19 +1,35 @@
module github.com/ClusterCockpit/cc-metric-collector
go 1.16
require (
github.com/NVIDIA/go-nvml v0.11.1-0
github.com/influxdata/influxdb-client-go/v2 v2.7.0
github.com/influxdata/line-protocol v0.0.0-20210922203350-b1ad95c89adf
github.com/nats-io/nats.go v1.13.1-0.20211122170419-d7c1d78a50fc
golang.org/x/sys v0.0.0-20220114195835-da31bd327af9
gopkg.in/Knetic/govaluate.v2 v2.3.0
)
go 1.17
require (
github.com/ClusterCockpit/cc-units v0.0.0-20220318130935-92a0c6442220
github.com/NVIDIA/go-nvml v0.11.6-0
github.com/PaesslerAG/gval v1.1.2
github.com/golang/protobuf v1.5.2 // indirect
github.com/nats-io/nats-server/v2 v2.7.0 // indirect
google.golang.org/protobuf v1.27.1 // indirect
github.com/gorilla/mux v1.8.0
github.com/influxdata/influxdb-client-go/v2 v2.8.1
github.com/influxdata/line-protocol v0.0.0-20210922203350-b1ad95c89adf
github.com/nats-io/nats.go v1.14.0
github.com/prometheus/client_golang v1.12.1
github.com/stmcginnis/gofish v0.13.0
golang.org/x/sys v0.0.0-20220412211240-33da011f77ad
)
require (
github.com/beorn7/perks v1.0.1 // indirect
github.com/cespare/xxhash/v2 v2.1.2 // indirect
github.com/deepmap/oapi-codegen v1.8.2 // indirect
github.com/golang/protobuf v1.5.2 // indirect
github.com/matttproud/golang_protobuf_extensions v1.0.1 // indirect
github.com/nats-io/nats-server/v2 v2.8.0 // indirect
github.com/nats-io/nkeys v0.3.0 // indirect
github.com/nats-io/nuid v1.0.1 // indirect
github.com/pkg/errors v0.9.1 // indirect
github.com/prometheus/client_model v0.2.0 // indirect
github.com/prometheus/common v0.32.1 // indirect
github.com/prometheus/procfs v0.7.3 // indirect
golang.org/x/crypto v0.0.0-20220112180741-5e0467b6c7ce // indirect
golang.org/x/net v0.0.0-20211112202133-69e39bad7dc2 // indirect
google.golang.org/protobuf v1.26.0 // indirect
gopkg.in/yaml.v2 v2.4.0 // indirect
)

16
go.mod.1.16 Normal file
View File

@@ -0,0 +1,16 @@
module github.com/ClusterCockpit/cc-metric-collector
go 1.16
require (
github.com/NVIDIA/go-nvml v0.11.6-0
github.com/PaesslerAG/gval v1.1.2
github.com/gorilla/mux v1.8.0
github.com/influxdata/influxdb-client-go/v2 v2.7.0
github.com/influxdata/line-protocol v0.0.0-20210922203350-b1ad95c89adf
github.com/nats-io/nats-server/v2 v2.8.0 // indirect
github.com/nats-io/nats.go v1.14.0
github.com/prometheus/client_golang v1.12.1
github.com/stmcginnis/gofish v0.13.0
golang.org/x/sys v0.0.0-20220412211240-33da011f77ad
)

16
go.mod.1.17+ Normal file
View File

@@ -0,0 +1,16 @@
module github.com/ClusterCockpit/cc-metric-collector
go 1.17
require (
github.com/NVIDIA/go-nvml v0.11.6-0
github.com/PaesslerAG/gval v1.1.2
github.com/gorilla/mux v1.8.0
github.com/influxdata/influxdb-client-go/v2 v2.8.1
github.com/influxdata/line-protocol v0.0.0-20210922203350-b1ad95c89adf
github.com/nats-io/nats-server/v2 v2.8.0 // indirect
github.com/nats-io/nats.go v1.14.0
github.com/prometheus/client_golang v1.12.1
github.com/stmcginnis/gofish v0.13.0
golang.org/x/sys v0.0.0-20220412211240-33da011f77ad
)

View File

@@ -24,8 +24,8 @@ type ccMetric struct {
// ccMetric access functions
type CCMetric interface {
ToPoint(metaAsTags bool) *write.Point // Generate influxDB point for data type ccMetric
ToLineProtocol(metaAsTags bool) string // Generate influxDB line protocol for data type ccMetric
ToPoint(metaAsTags map[string]bool) *write.Point // Generate influxDB point for data type ccMetric
ToLineProtocol(metaAsTags map[string]bool) string // Generate influxDB line protocol for data type ccMetric
Name() string // Get metric name
SetName(name string) // Set metric name
@@ -61,25 +61,18 @@ func (m *ccMetric) String() string {
}
// ToLineProtocol generates influxDB line protocol for data type ccMetric
func (m *ccMetric) ToPoint(metaAsTags bool) (p *write.Point) {
if !metaAsTags {
p = influxdb2.NewPoint(m.name, m.tags, m.fields, m.tm)
} else {
tags := make(map[string]string, len(m.tags)+len(m.meta))
for key, value := range m.tags {
tags[key] = value
func (m *ccMetric) ToPoint(metaAsTags map[string]bool) (p *write.Point) {
p = influxdb2.NewPoint(m.name, m.tags, m.fields, m.tm)
for key, ok1 := range metaAsTags {
if val, ok2 := m.GetMeta(key); ok1 && ok2 {
p.AddTag(key, val)
}
for key, value := range m.meta {
tags[key] = value
}
p = influxdb2.NewPoint(m.name, tags, m.fields, m.tm)
}
return
return p
}
// ToLineProtocol generates influxDB line protocol for data type ccMetric
func (m *ccMetric) ToLineProtocol(metaAsTags bool) string {
func (m *ccMetric) ToLineProtocol(metaAsTags map[string]bool) string {
return write.PointToLineProtocol(
m.ToPoint(metaAsTags),

View File

@@ -29,6 +29,7 @@ func intArrayContains(array []int, str int) (int, bool) {
return -1, false
}
// Used internally for sysfs file reads
func fileToInt(path string) int {
buffer, err := ioutil.ReadFile(path)
if err != nil {
@@ -47,6 +48,7 @@ func fileToInt(path string) int {
return int(id)
}
// Get list of CPU socket IDs
func SocketList() []int {
buffer, err := ioutil.ReadFile(string(PROCFS_CPUINFO))
if err != nil {
@@ -54,7 +56,7 @@ func SocketList() []int {
return nil
}
ll := strings.Split(string(buffer), "\n")
var packs []int
packs := make([]int, 0)
for _, line := range ll {
if strings.HasPrefix(line, "physical id") {
lv := strings.Fields(line)
@@ -72,7 +74,8 @@ func SocketList() []int {
return packs
}
func CpuList() []int {
// Get list of hardware thread IDs in the order of listing in /proc/cpuinfo
func HwthreadList() []int {
buffer, err := ioutil.ReadFile(string(PROCFS_CPUINFO))
if err != nil {
log.Print(err)
@@ -97,6 +100,13 @@ func CpuList() []int {
return cpulist
}
// Get list of hardware thread IDs in the order of listing in /proc/cpuinfo
// Deprecated! Use HwthreadList()
func CpuList() []int {
return HwthreadList()
}
// Get list of CPU core IDs in the order of listing in /proc/cpuinfo
func CoreList() []int {
buffer, err := ioutil.ReadFile(string(PROCFS_CPUINFO))
if err != nil {
@@ -122,6 +132,7 @@ func CoreList() []int {
return corelist
}
// Get list of NUMA node IDs
func NumaNodeList() []int {
numaList := make([]int, 0)
globPath := filepath.Join(string(SYSFS_NUMABASE), "node*")
@@ -156,8 +167,9 @@ func NumaNodeList() []int {
return numaList
}
// Get list of CPU die IDs
func DieList() []int {
cpulist := CpuList()
cpulist := HwthreadList()
dielist := make([]int, 0)
for _, c := range cpulist {
diepath := filepath.Join(string(SYSFS_CPUBASE), fmt.Sprintf("cpu%d", c), "topology/die_id")
@@ -169,10 +181,33 @@ func DieList() []int {
}
}
}
return dielist
if len(dielist) > 0 {
return dielist
}
return SocketList()
}
type CpuEntry struct {
// Get list of specified type using the naming format inside ClusterCockpit
func GetTypeList(topology_type string) []int {
switch topology_type {
case "node":
return []int{0}
case "socket":
return SocketList()
case "die":
return DieList()
case "memoryDomain":
return NumaNodeList()
case "core":
return CoreList()
case "hwthread":
return HwthreadList()
}
return []int{}
}
// Structure holding all information about a hardware thread
type HwthreadEntry struct {
Cpuid int
SMT int
Core int
@@ -181,25 +216,25 @@ type CpuEntry struct {
Die int
}
func CpuData() []CpuEntry {
func CpuData() []HwthreadEntry {
fileToInt := func(path string) int {
buffer, err := ioutil.ReadFile(path)
if err != nil {
log.Print(err)
//cclogger.ComponentError("ccTopology", "Reading", path, ":", err.Error())
return -1
}
sbuffer := strings.Replace(string(buffer), "\n", "", -1)
var id int64
//_, err = fmt.Scanf("%d", sbuffer, &id)
id, err = strconv.ParseInt(sbuffer, 10, 32)
if err != nil {
cclogger.ComponentError("ccTopology", "Parsing", path, ":", sbuffer, err.Error())
return -1
}
return int(id)
}
// fileToInt := func(path string) int {
// buffer, err := ioutil.ReadFile(path)
// if err != nil {
// log.Print(err)
// //cclogger.ComponentError("ccTopology", "Reading", path, ":", err.Error())
// return -1
// }
// sbuffer := strings.Replace(string(buffer), "\n", "", -1)
// var id int64
// //_, err = fmt.Scanf("%d", sbuffer, &id)
// id, err = strconv.ParseInt(sbuffer, 10, 32)
// if err != nil {
// cclogger.ComponentError("ccTopology", "Parsing", path, ":", sbuffer, err.Error())
// return -1
// }
// return int(id)
// }
getCore := func(basepath string) int {
return fileToInt(fmt.Sprintf("%s/core_id", basepath))
}
@@ -257,11 +292,11 @@ func CpuData() []CpuEntry {
return 0
}
clist := make([]CpuEntry, 0)
for _, c := range CpuList() {
clist = append(clist, CpuEntry{Cpuid: c})
clist := make([]HwthreadEntry, 0)
for _, c := range HwthreadList() {
clist = append(clist, HwthreadEntry{Cpuid: c})
}
for _, centry := range clist {
for i, centry := range clist {
centry.Socket = -1
centry.Numadomain = -1
centry.Die = -1
@@ -289,10 +324,13 @@ func CpuData() []CpuEntry {
// Lookup NUMA domain id
centry.Numadomain = getNumaDomain(base)
// Update values in output list
clist[i] = centry
}
return clist
}
// Structure holding basic information about a CPU
type CpuInformation struct {
NumHWthreads int
SMTWidth int
@@ -302,6 +340,7 @@ type CpuInformation struct {
NumNumaDomains int
}
// Get basic information about the CPU
func CpuInfo() CpuInformation {
var c CpuInformation
@@ -337,7 +376,8 @@ func CpuInfo() CpuInformation {
return c
}
func GetCpuSocket(cpuid int) int {
// Get the CPU socket ID for a given hardware thread ID
func GetHwthreadSocket(cpuid int) int {
cdata := CpuData()
for _, d := range cdata {
if d.Cpuid == cpuid {
@@ -347,7 +387,8 @@ func GetCpuSocket(cpuid int) int {
return -1
}
func GetCpuNumaDomain(cpuid int) int {
// Get the NUMA node ID for a given hardware thread ID
func GetHwthreadNumaDomain(cpuid int) int {
cdata := CpuData()
for _, d := range cdata {
if d.Cpuid == cpuid {
@@ -357,7 +398,8 @@ func GetCpuNumaDomain(cpuid int) int {
return -1
}
func GetCpuDie(cpuid int) int {
// Get the CPU die ID for a given hardware thread ID
func GetHwthreadDie(cpuid int) int {
cdata := CpuData()
for _, d := range cdata {
if d.Cpuid == cpuid {
@@ -367,7 +409,8 @@ func GetCpuDie(cpuid int) int {
return -1
}
func GetCpuCore(cpuid int) int {
// Get the CPU core ID for a given hardware thread ID
func GetHwthreadCore(cpuid int) int {
cdata := CpuData()
for _, d := range cdata {
if d.Cpuid == cpuid {
@@ -377,7 +420,8 @@ func GetCpuCore(cpuid int) int {
return -1
}
func GetSocketCpus(socket int) []int {
// Get the all hardware thread ID associated with a CPU socket
func GetSocketHwthreads(socket int) []int {
all := CpuData()
cpulist := make([]int, 0)
for _, d := range all {
@@ -388,7 +432,8 @@ func GetSocketCpus(socket int) []int {
return cpulist
}
func GetNumaDomainCpus(domain int) []int {
// Get the all hardware thread ID associated with a NUMA node
func GetNumaDomainHwthreads(domain int) []int {
all := CpuData()
cpulist := make([]int, 0)
for _, d := range all {
@@ -399,7 +444,8 @@ func GetNumaDomainCpus(domain int) []int {
return cpulist
}
func GetDieCpus(die int) []int {
// Get the all hardware thread ID associated with a CPU die
func GetDieHwthreads(die int) []int {
all := CpuData()
cpulist := make([]int, 0)
for _, d := range all {
@@ -410,7 +456,8 @@ func GetDieCpus(die int) []int {
return cpulist
}
func GetCoreCpus(core int) []int {
// Get the all hardware thread ID associated with a CPU core
func GetCoreHwthreads(core int) []int {
all := CpuData()
cpulist := make([]int, 0)
for _, d := range all {

View File

@@ -246,7 +246,7 @@ func matchfunc(args ...interface{}) (interface{}, error) {
func getCpuCoreFunc(args ...interface{}) (interface{}, error) {
switch cpuid := args[0].(type) {
case int:
return topo.GetCpuCore(cpuid), nil
return topo.GetHwthreadCore(cpuid), nil
}
return -1, errors.New("function 'getCpuCore' accepts only an 'int' cpuid")
}
@@ -255,7 +255,7 @@ func getCpuCoreFunc(args ...interface{}) (interface{}, error) {
func getCpuSocketFunc(args ...interface{}) (interface{}, error) {
switch cpuid := args[0].(type) {
case int:
return topo.GetCpuSocket(cpuid), nil
return topo.GetHwthreadSocket(cpuid), nil
}
return -1, errors.New("function 'getCpuCore' accepts only an 'int' cpuid")
}
@@ -264,7 +264,7 @@ func getCpuSocketFunc(args ...interface{}) (interface{}, error) {
func getCpuNumaDomainFunc(args ...interface{}) (interface{}, error) {
switch cpuid := args[0].(type) {
case int:
return topo.GetCpuNumaDomain(cpuid), nil
return topo.GetHwthreadNumaDomain(cpuid), nil
}
return -1, errors.New("function 'getCpuNuma' accepts only an 'int' cpuid")
}
@@ -273,7 +273,7 @@ func getCpuNumaDomainFunc(args ...interface{}) (interface{}, error) {
func getCpuDieFunc(args ...interface{}) (interface{}, error) {
switch cpuid := args[0].(type) {
case int:
return topo.GetCpuDie(cpuid), nil
return topo.GetHwthreadDie(cpuid), nil
}
return -1, errors.New("function 'getCpuDie' accepts only an 'int' cpuid")
}
@@ -336,7 +336,7 @@ func getCpuListOfDieFunc(args ...interface{}) (interface{}, error) {
// wrapper function to get a list of all cpuids of the node
func getCpuListOfNode(args ...interface{}) (interface{}, error) {
return topo.CpuList(), nil
return topo.HwthreadList(), nil
}
// helper function to get the cpuid list for a CCMetric type tag set (type and type-id)
@@ -348,14 +348,14 @@ func getCpuListOfType(args ...interface{}) (interface{}, error) {
case string:
switch typ {
case "node":
return topo.CpuList(), nil
return topo.HwthreadList(), nil
case "socket":
return getCpuListOfSocketFunc(args[1])
case "numadomain":
return getCpuListOfNumaDomainFunc(args[1])
case "core":
return getCpuListOfCoreFunc(args[1])
case "cpu":
case "hwthread":
var cpu int
switch id := args[1].(type) {

View File

@@ -8,6 +8,8 @@ The CCMetric router sits in between the collectors and the sinks and can be used
{
"num_cache_intervals" : 1,
"interval_timestamp" : true,
"hostname_tag" : "hostname",
"max_forward" : 50,
"add_tags" : [
{
"key" : "cluster",
@@ -50,11 +52,30 @@ The CCMetric router sits in between the collectors and the sinks and can be used
],
"rename_metrics" : {
"metric_12345" : "mymetric"
},
"normalize_units" : true,
"change_unit_prefix" : {
"mem_used" : "G",
"mem_total" : "G"
}
}
```
There are three main options `add_tags`, `delete_tags` and `interval_timestamp`. `add_tags` and `delete_tags` are lists consisting of dicts with `key`, `value` and `if`. The `value` can be omitted in the `delete_tags` part as it only uses the `key` for removal. The `interval_timestamp` setting means that a unique timestamp is applied to all metrics traversing the router during an interval.
# Processing order in the router
- Add the `hostname_tag` tag (if sent by collectors or cache)
- If `interval_timestamp == true`, change time of metrics
- Check if metric should be dropped (`drop_metrics` and `drop_metrics_if`)
- Add tags from `add_tags`
- Delete tags from `del_tags`
- Rename metric based on `rename_metrics` and store old name as `oldname` in meta information
- Add tags from `add_tags` (if you used the new name in the `if` condition)
- Delete tags from `del_tags` (if you used the new name in the `if` condition)
- Send to sinks
- Move to cache (if `num_cache_intervals > 0`)
# The `interval_timestamp` option
The collectors' `Read()` functions are not called simultaneously and therefore the metrics gathered in an interval can have different timestamps. If you want to avoid that and have a common timestamp (the beginning of the interval), set this option to `true` and the MetricRouter sets the time.
@@ -65,6 +86,14 @@ If the MetricRouter should buffer metrics of intervals in a MetricCache, this op
A `num_cache_intervals > 0` is required to use the `interval_aggregates` option.
# The `hostname_tag` option
By default, the router tags metrics with the hostname for all locally created metrics. The default tag name is `hostname`, but it can be changed if your organization wants anything else
# The `max_forward` option
Every time the router receives a metric through any of the channels, it tries to directly read up to `max_forward` metrics from the same channel. This was done as the router thread would go to sleep and wake up with every arriving metric. The default are `50` metrics at once and `max_forward` needs to greater than `1`.
# The `rename_metrics` option
In the ClusterCockpit world we specified a set of standard metrics. Since some collectors determine the metric names based on files, execuables and libraries, they might change from system to system (or installation to installtion, OS to OS, ...). In order to get the common names, you can rename incoming metrics before sending them to the sink. If the metric name matches the `oldname`, it is changed to `newname`
@@ -168,6 +197,14 @@ This option takes a list of evaluable conditions and performs them one after the
```
The first line is comparable with the example in `drop_metrics`, it drops all metrics starting with `drop_metric_` and ending with a number. The second line drops all metrics of the first hardware thread (**not** recommended)
# Manipulating the metric units
## The `normalize_units` option
The cc-metric-collector tries to read the data from the system as it is reported. If available, it tries to read the metric unit from the system as well (e.g. from `/proc/meminfo`). The problem is that, depending on the source, the metric units are named differently. Just think about `byte`, `Byte`, `B`, `bytes`, ...
The [cc-units](https://github.com/ClusterCockpit/cc-units) package provides us a normalization option to use the same metric unit name for all metrics. It this option is set to true, all `unit` meta tags are normalized.
## The `change_unit_prefix` section
It is often the case that metrics are reported by the system using a rather outdated unit prefix (like `/proc/meminfo` still uses kByte despite current memory sizes are in the GByte range). If you want to change the prefix of a unit, you can do that with the help of [cc-units](https://github.com/ClusterCockpit/cc-units). The setting works on the metric name and requires the new prefix for the metric. The cc-units package determines the scaling factor.
# Aggregate metric values of the current interval with the `interval_aggregates` option
@@ -215,3 +252,22 @@ Use cases for `interval_aggregates`:
}
}
```
# Order of operations
The router performs the above mentioned options in a specific order. In order to get the logic you want for a specific metric, it is crucial to know the processing order:
- Add the `hostname` tag (c)
- Manipulate the timestamp to the interval timestamp (c,r)
- Drop metrics based on `drop_metrics` and `drop_metrics_if` (c,r)
- Add tags based on `add_tags` (c,r)
- Delete tags based on `del_tags` (c,r)
- Rename metric based on `rename_metric` (c,r)
- Add tags based on `add_tags` to still work if the configuration uses the new name (c,r)
- Delete tags based on `del_tags` to still work if the configuration uses the new name (c,r)
- Normalize units when `normalize_units` is set (c,r)
- Convert unit prefix based on `change_unit_prefix` (c,r)
Legend:
- 'c' if metric is coming from a collector
- 'r' if metric is coming from a receiver

View File

@@ -12,6 +12,7 @@ import (
lp "github.com/ClusterCockpit/cc-metric-collector/internal/ccMetric"
agg "github.com/ClusterCockpit/cc-metric-collector/internal/metricAggregator"
mct "github.com/ClusterCockpit/cc-metric-collector/internal/multiChanTicker"
units "github.com/ClusterCockpit/cc-units"
)
const ROUTER_MAX_FORWARD = 50
@@ -25,6 +26,7 @@ type metricRouterTagConfig struct {
// Metric router configuration
type metricRouterConfig struct {
HostnameTagName string `json:"hostname_tag"` // Key name used when adding the hostname to a metric (default 'hostname')
AddTags []metricRouterTagConfig `json:"add_tags"` // List of tags that are added when the condition is met
DelTags []metricRouterTagConfig `json:"delete_tags"` // List of tags that are removed when the condition is met
IntervalAgg []agg.MetricAggregatorIntervalConfig `json:"interval_aggregates"` // List of aggregation function processed at the end of an interval
@@ -33,6 +35,9 @@ type metricRouterConfig struct {
RenameMetrics map[string]string `json:"rename_metrics"` // Map to rename metric name from key to value
IntervalStamp bool `json:"interval_timestamp"` // Update timestamp periodically by ticker each interval?
NumCacheIntervals int `json:"num_cache_intervals"` // Number of intervals of cached metrics for evaluation
MaxForward int `json:"max_forward"` // Number of maximal forwarded metrics at one select
NormalizeUnits bool `json:"normalize_units"` // Check unit meta flag and normalize it using cc-units
ChangeUnitPrefix map[string]string `json:"change_unit_prefix"` // Add prefix that should be applied to the metrics
dropMetrics map[string]bool // Internal map for O(1) lookup
}
@@ -46,7 +51,6 @@ type metricRouter struct {
done chan bool // channel to finish / stop metric router
wg *sync.WaitGroup // wait group for all goroutines in cc-metric-collector
timestamp time.Time // timestamp periodically updated by ticker each interval
timerdone chan bool // channel to finish / stop timestamp updater
ticker mct.MultiChanTicker // periodically ticking once each interval
config metricRouterConfig // json encoded config for metric router
cache MetricCache // pointer to MetricCache
@@ -76,7 +80,8 @@ func (r *metricRouter) Init(ticker mct.MultiChanTicker, wg *sync.WaitGroup, rout
r.cache_input = make(chan lp.CCMetric)
r.wg = wg
r.ticker = ticker
r.maxForward = ROUTER_MAX_FORWARD
r.config.MaxForward = ROUTER_MAX_FORWARD
r.config.HostnameTagName = "hostname"
// Set hostname
hostname, err := os.Hostname()
@@ -100,6 +105,10 @@ func (r *metricRouter) Init(ticker mct.MultiChanTicker, wg *sync.WaitGroup, rout
cclog.ComponentError("MetricRouter", err.Error())
return err
}
r.maxForward = 1
if r.config.MaxForward > r.maxForward {
r.maxForward = r.config.MaxForward
}
if r.config.NumCacheIntervals > 0 {
r.cache, err = NewCache(r.cache_input, r.ticker, &r.cachewg, r.config.NumCacheIntervals)
if err != nil {
@@ -117,29 +126,6 @@ func (r *metricRouter) Init(ticker mct.MultiChanTicker, wg *sync.WaitGroup, rout
return nil
}
// StartTimer starts a timer which updates timestamp periodically
func (r *metricRouter) StartTimer() {
m := make(chan time.Time)
r.ticker.AddChannel(m)
r.timerdone = make(chan bool)
r.wg.Add(1)
go func() {
defer r.wg.Done()
for {
select {
case <-r.timerdone:
close(r.timerdone)
cclog.ComponentDebug("MetricRouter", "TIMER DONE")
return
case t := <-m:
r.timestamp = t
}
}
}()
cclog.ComponentDebug("MetricRouter", "TIMER START")
}
func getParamMap(point lp.CCMetric) map[string]interface{} {
params := make(map[string]interface{})
params["metric"] = point
@@ -224,12 +210,45 @@ func (r *metricRouter) dropMetric(point lp.CCMetric) bool {
return false
}
func (r *metricRouter) prepareUnit(point lp.CCMetric) bool {
if r.config.NormalizeUnits {
if in_unit, ok := point.GetMeta("unit"); ok {
u := units.NewUnit(in_unit)
if u.Valid() {
point.AddMeta("unit", u.Short())
}
}
}
if newP, ok := r.config.ChangeUnitPrefix[point.Name()]; ok {
newPrefix := units.NewPrefix(newP)
if in_unit, ok := point.GetMeta("unit"); ok && newPrefix != units.InvalidPrefix {
u := units.NewUnit(in_unit)
if u.Valid() {
cclog.ComponentDebug("MetricRouter", "Change prefix to", newP, "for metric", point.Name())
conv, out_unit := units.GetUnitPrefixFactor(u, newPrefix)
if conv != nil && out_unit.Valid() {
if val, ok := point.GetField("value"); ok {
point.AddField("value", conv(val))
point.AddMeta("unit", out_unit.Short())
}
}
}
}
}
return true
}
// Start starts the metric router
func (r *metricRouter) Start() {
// start timer if configured
r.timestamp = time.Now()
timeChan := make(chan time.Time)
if r.config.IntervalStamp {
r.StartTimer()
r.ticker.AddChannel(timeChan)
}
// Router manager is done
@@ -244,11 +263,15 @@ func (r *metricRouter) Start() {
cclog.ComponentDebug("MetricRouter", "FORWARD", point)
r.DoAddTags(point)
r.DoDelTags(point)
if new, ok := r.config.RenameMetrics[point.Name()]; ok {
name := point.Name()
if new, ok := r.config.RenameMetrics[name]; ok {
point.SetName(new)
point.AddMeta("oldname", name)
r.DoAddTags(point)
r.DoDelTags(point)
}
r.DoAddTags(point)
r.DoDelTags(point)
r.prepareUnit(point)
for _, o := range r.outputs {
o <- point
@@ -258,7 +281,7 @@ func (r *metricRouter) Start() {
// Foward message received from collector channel
coll_forward := func(p lp.CCMetric) {
// receive from metric collector
p.AddTag("hostname", r.hostname)
p.AddTag(r.config.HostnameTagName, r.hostname)
if r.config.IntervalStamp {
p.SetTime(r.timestamp)
}
@@ -287,7 +310,7 @@ func (r *metricRouter) Start() {
cache_forward := func(p lp.CCMetric) {
// receive from metric collector
if !r.dropMetric(p) {
p.AddTag("hostname", r.hostname)
p.AddTag(r.config.HostnameTagName, r.hostname)
forward(p)
}
}
@@ -307,21 +330,25 @@ func (r *metricRouter) Start() {
done()
return
case timestamp := <-timeChan:
r.timestamp = timestamp
cclog.ComponentDebug("MetricRouter", "Update timestamp", r.timestamp.UnixNano())
case p := <-r.coll_input:
coll_forward(p)
for i := 0; len(r.coll_input) > 0 && i < r.maxForward; i++ {
for i := 0; len(r.coll_input) > 0 && i < (r.maxForward-1); i++ {
coll_forward(<-r.coll_input)
}
case p := <-r.recv_input:
recv_forward(p)
for i := 0; len(r.recv_input) > 0 && i < r.maxForward; i++ {
for i := 0; len(r.recv_input) > 0 && i < (r.maxForward-1); i++ {
recv_forward(<-r.recv_input)
}
case p := <-r.cache_input:
cache_forward(p)
for i := 0; len(r.cache_input) > 0 && i < r.maxForward; i++ {
for i := 0; len(r.cache_input) > 0 && i < (r.maxForward-1); i++ {
cache_forward(<-r.cache_input)
}
}
@@ -352,14 +379,6 @@ func (r *metricRouter) Close() {
// wait for close of channel r.done
<-r.done
// stop timer
if r.config.IntervalStamp {
cclog.ComponentDebug("MetricRouter", "TIMER CLOSE")
r.timerdone <- true
// wait for close of channel r.timerdone
<-r.timerdone
}
// stop metric cache
if r.config.NumCacheIntervals > 0 {
cclog.ComponentDebug("MetricRouter", "CACHE CLOSE")

View File

@@ -4,5 +4,22 @@
"address": "nats://my-url",
"port" : "4222",
"database": "testcluster"
},
"redfish_recv": {
"type": "redfish",
"client_config": [
{
"hostname": "my-host-1",
"username": "username-1",
"password": "password-1",
"endpoint": "https://my-endpoint-1"
},
{
"hostname": "my-host-2",
"username": "username-2",
"password": "password-2",
"endpoint": "https://my-endpoint-2"
}
]
}
}

View File

@@ -21,6 +21,7 @@ This allows to specify
- [`nats`](./natsReceiver.md): Receive metrics from the NATS network
- [`prometheus`](./prometheusReceiver.md): Scrape data from a Prometheus client
- [`http`](./httpReceiver.md): Listen for HTTP Post requests transporting metrics in InfluxDB line protocol
# Contributing own receivers
A receiver contains a few functions and is derived from the type `Receiver` (in `metricReceiver.go`):

118
receivers/httpReceiver.go Normal file
View File

@@ -0,0 +1,118 @@
package receivers
import (
"context"
"encoding/json"
"errors"
"fmt"
"io/ioutil"
"net/http"
"strings"
"sync"
cclog "github.com/ClusterCockpit/cc-metric-collector/internal/ccLogger"
lp "github.com/ClusterCockpit/cc-metric-collector/internal/ccMetric"
"github.com/gorilla/mux"
influx "github.com/influxdata/line-protocol"
)
const HTTP_RECEIVER_PORT = "8080"
type HttpReceiverConfig struct {
Type string `json:"type"`
Addr string `json:"address"`
Port string `json:"port"`
Path string `json:"path"`
}
type HttpReceiver struct {
receiver
handler *influx.MetricHandler
parser *influx.Parser
meta map[string]string
config HttpReceiverConfig
router *mux.Router
server *http.Server
wg sync.WaitGroup
}
func (r *HttpReceiver) Init(name string, config json.RawMessage) error {
r.name = fmt.Sprintf("HttpReceiver(%s)", name)
r.config.Port = HTTP_RECEIVER_PORT
if len(config) > 0 {
err := json.Unmarshal(config, &r.config)
if err != nil {
cclog.ComponentError(r.name, "Error reading config:", err.Error())
return err
}
}
if len(r.config.Port) == 0 {
return errors.New("not all configuration variables set required by HttpReceiver")
}
r.meta = map[string]string{"source": r.name}
p := r.config.Path
if !strings.HasPrefix(p, "/") {
p = "/" + p
}
uri := fmt.Sprintf("%s:%s%s", r.config.Addr, r.config.Port, p)
cclog.ComponentDebug(r.name, "INIT", uri)
r.handler = influx.NewMetricHandler()
r.parser = influx.NewParser(r.handler)
r.parser.SetTimeFunc(DefaultTime)
r.router = mux.NewRouter()
r.router.Path(p).HandlerFunc(r.ServerHttp)
r.server = &http.Server{Addr: uri, Handler: r.router}
return nil
}
func (r *HttpReceiver) Start() {
cclog.ComponentDebug(r.name, "START")
r.wg.Add(1)
go func() {
err := r.server.ListenAndServe()
if err != nil && err.Error() != "http: Server closed" {
cclog.ComponentError(r.name, err.Error())
}
r.wg.Done()
}()
}
func (r *HttpReceiver) ServerHttp(w http.ResponseWriter, req *http.Request) {
if req.Method != http.MethodPost {
http.Error(w, "Method Not Allowed", http.StatusMethodNotAllowed)
return
}
body, err := ioutil.ReadAll(req.Body)
if err != nil {
http.Error(w, err.Error(), http.StatusInternalServerError)
return
}
metrics, err := r.parser.Parse(body)
if err != nil {
http.Error(w, err.Error(), http.StatusBadRequest)
}
for _, m := range metrics {
y := lp.FromInfluxMetric(m)
for k, v := range r.meta {
y.AddMeta(k, v)
}
if r.sink != nil {
r.sink <- y
}
}
w.WriteHeader(http.StatusOK)
}
func (r *HttpReceiver) Close() {
r.server.Shutdown(context.Background())
}
func NewHttpReceiver(name string, config json.RawMessage) (Receiver, error) {
r := new(HttpReceiver)
err := r.Init(name, config)
return r, err
}

23
receivers/httpReceiver.md Normal file
View File

@@ -0,0 +1,23 @@
## `http` receiver
The `http` receiver can be used receive metrics through HTTP POST requests.
### Configuration structure
```json
{
"<name>": {
"type": "http",
"address" : "",
"port" : "8080",
"path" : "/write"
}
}
```
- `type`: makes the receiver a `http` receiver
- `address`: Listen address
- `port`: Listen port
- `path`: URL path for the write endpoint
The HTTP endpoint listens to `http://<address>:<port>/<path>`

View File

@@ -10,14 +10,13 @@ import (
)
var AvailableReceivers = map[string]func(name string, config json.RawMessage) (Receiver, error){
"nats": NewNatsReceiver,
"nats": NewNatsReceiver,
"redfish": NewRedfishReceiver,
}
type receiveManager struct {
inputs []Receiver
output chan lp.CCMetric
done chan bool
wg *sync.WaitGroup
config []json.RawMessage
}
@@ -33,8 +32,6 @@ func (rm *receiveManager) Init(wg *sync.WaitGroup, receiverConfigFile string) er
// Initialize struct fields
rm.inputs = make([]Receiver, 0)
rm.output = nil
rm.done = make(chan bool)
rm.wg = wg
rm.config = make([]json.RawMessage, 0)
configFile, err := os.Open(receiverConfigFile)
@@ -58,7 +55,7 @@ func (rm *receiveManager) Init(wg *sync.WaitGroup, receiverConfigFile string) er
}
func (rm *receiveManager) Start() {
rm.wg.Add(1)
cclog.ComponentDebug("ReceiveManager", "START")
for _, r := range rm.inputs {
cclog.ComponentDebug("ReceiveManager", "START", r.Name())
@@ -97,16 +94,19 @@ func (rm *receiveManager) AddOutput(output chan lp.CCMetric) {
}
func (rm *receiveManager) Close() {
cclog.ComponentDebug("ReceiveManager", "CLOSE")
// Close all receivers
for _, r := range rm.inputs {
cclog.ComponentDebug("ReceiveManager", "CLOSE", r.Name())
r.Close()
}
rm.wg.Done()
cclog.ComponentDebug("ReceiveManager", "CLOSE")
cclog.ComponentDebug("ReceiveManager", "DONE")
}
func New(wg *sync.WaitGroup, receiverConfigFile string) (ReceiveManager, error) {
r := &receiveManager{}
r := new(receiveManager)
err := r.Init(wg, receiverConfigFile)
if err != nil {
return nil, err

View File

@@ -0,0 +1,324 @@
package receivers
import (
"encoding/json"
"fmt"
"strconv"
"sync"
"time"
cclog "github.com/ClusterCockpit/cc-metric-collector/internal/ccLogger"
lp "github.com/ClusterCockpit/cc-metric-collector/internal/ccMetric"
// See: https://pkg.go.dev/github.com/stmcginnis/gofish
"github.com/stmcginnis/gofish"
)
// RedfishReceiver configuration:
type RedfishReceiver struct {
receiver
config struct {
Type string `json:"type"`
Fanout int `json:"fanout,omitempty"` // Default fanout: 64
Interval int `json:"interval,omitempty"` // Default interval: 30s
// Client config for each redfish service
ClientConfigs []struct {
Hostname *string `json:"hostname"`
Username *string `json:"username"`
Password *string `json:"password"`
Endpoint *string `json:"endpoint"`
Insecure *bool `json:"insecure,omitempty"`
ExcludeMetrics []string `json:"exclude_metrics,omitempty"`
gofish gofish.ClientConfig
} `json:"client_config"`
}
done chan bool // channel to finish / stop redfish receiver
wg sync.WaitGroup // wait group for redfish receiver
}
// Start starts the redfish receiver
func (r *RedfishReceiver) Start() {
cclog.ComponentDebug(r.name, "START")
// readPowerMetric reads readfish power metric from the endpoint configured in conf
readPowerMetric := func(clientConfigIndex int) error {
clientConfig := &r.config.ClientConfigs[clientConfigIndex]
// Connect to redfish service
c, err := gofish.Connect(clientConfig.gofish)
if err != nil {
c := struct {
Username string
Endpoint string
BasicAuth bool
Insecure bool
}{
Username: clientConfig.gofish.Username,
Endpoint: clientConfig.gofish.Endpoint,
BasicAuth: clientConfig.gofish.BasicAuth,
Insecure: clientConfig.gofish.Insecure,
}
return fmt.Errorf("readPowerMetric: gofish.Connect(%+v) failed: %v", c, err)
}
defer c.Logout()
// Get all chassis managed by this service
chassis_list, err := c.Service.Chassis()
if err != nil {
return fmt.Errorf("readPowerMetric: c.Service.Chassis() failed: %v", err)
}
for _, chassis := range chassis_list {
timestamp := time.Now()
// Get power information for each chassis
power, err := chassis.Power()
if err != nil {
return fmt.Errorf("readPowerMetric: chassis.Power() failed: %v", err)
}
if power == nil {
continue
}
// Read min, max and average consumed watts for each power control
for _, pc := range power.PowerControl {
// Map of collected metrics
metrics := map[string]float32{
// PowerConsumedWatts shall represent the actual power being consumed (in
// Watts) by the chassis
"consumed_watts": pc.PowerConsumedWatts,
// AverageConsumedWatts shall represent the
// average power level that occurred averaged over the last IntervalInMin
// minutes.
"average_consumed_watts": pc.PowerMetrics.AverageConsumedWatts,
// MinConsumedWatts shall represent the
// minimum power level in watts that occurred within the last
// IntervalInMin minutes.
"min_consumed_watts": pc.PowerMetrics.MinConsumedWatts,
// MaxConsumedWatts shall represent the
// maximum power level in watts that occurred within the last
// IntervalInMin minutes
"max_consumed_watts": pc.PowerMetrics.MaxConsumedWatts,
}
intervalInMin := strconv.FormatFloat(float64(pc.PowerMetrics.IntervalInMin), 'f', -1, 32)
// Metrics to exclude
for _, key := range clientConfig.ExcludeMetrics {
delete(metrics, key)
}
// Set tags
tags := map[string]string{
"hostname": *clientConfig.Hostname,
"type": "node",
// ID uniquely identifies the resource
"id": pc.ID,
// MemberID shall uniquely identify the member within the collection. For
// services supporting Redfish v1.6 or higher, this value shall be the
// zero-based array index.
"member_id": pc.MemberID,
// PhysicalContext shall be a description of the affected device(s) or region
// within the chassis to which this power control applies.
"physical_context": string(pc.PhysicalContext),
// Name
"power_control_name": pc.Name,
}
// Delete empty tags
for key, value := range tags {
if value == "" {
delete(tags, key)
}
}
// Set meta data tags
meta := map[string]string{
"source": r.name,
"group": "Energy",
"interval_in_minutes": intervalInMin,
"unit": "watts",
}
// Delete empty meta data tags
for key, value := range meta {
if value == "" {
delete(meta, key)
}
}
for name, value := range metrics {
y, err := lp.New(name, tags, meta,
map[string]interface{}{
"value": value,
},
timestamp)
if err == nil {
r.sink <- y
}
}
}
}
return nil
}
// doReadPowerMetric read power metrics for all configure redfish services.
// To compensate latencies of the Redfish services a fanout is used.
doReadPowerMetric := func() {
// Compute fanout to use
realFanout := r.config.Fanout
if len(r.config.ClientConfigs) < realFanout {
realFanout = len(r.config.ClientConfigs)
}
// Create wait group and input channel for workers
var workerWaitGroup sync.WaitGroup
workerInput := make(chan int, realFanout)
// Create worker go routines
for i := 0; i < realFanout; i++ {
// Increment worker wait group counter
workerWaitGroup.Add(1)
go func() {
// Decrement worker wait group counter
defer workerWaitGroup.Done()
// Read power metrics for each client config
for clientConfigIndex := range workerInput {
err := readPowerMetric(clientConfigIndex)
if err != nil {
cclog.ComponentError(r.name, err)
}
}
}()
}
// Distribute client configs to workers
for i := range r.config.ClientConfigs {
// Check done channel status
select {
case workerInput <- i:
case <-r.done:
// process done event
// Stop workers, clear channel and wait for all workers to finish
close(workerInput)
for range workerInput {
}
workerWaitGroup.Wait()
return
}
}
// Stop workers and wait for all workers to finish
close(workerInput)
workerWaitGroup.Wait()
}
// Start redfish receiver
r.wg.Add(1)
go func() {
defer r.wg.Done()
// Create ticker
ticker := time.NewTicker(time.Duration(r.config.Interval) * time.Second)
defer ticker.Stop()
for {
doReadPowerMetric()
select {
case <-ticker.C:
// process ticker event -> continue
continue
case <-r.done:
// process done event
return
}
}
}()
cclog.ComponentDebug(r.name, "STARTED")
}
// Close redfish receiver
func (r *RedfishReceiver) Close() {
cclog.ComponentDebug(r.name, "CLOSE")
// Send the signal and wait
close(r.done)
r.wg.Wait()
cclog.ComponentDebug(r.name, "DONE")
}
// New function to create a new instance of the receiver
// Initialize the receiver by giving it a name and reading in the config JSON
func NewRedfishReceiver(name string, config json.RawMessage) (Receiver, error) {
r := new(RedfishReceiver)
// Set name
r.name = fmt.Sprintf("RedfishReceiver(%s)", name)
// Create done channel
r.done = make(chan bool)
// Set defaults in r.config
// Allow overwriting these defaults by reading config JSON
r.config.Fanout = 64
r.config.Interval = 30
// Read the redfish receiver specific JSON config
if len(config) > 0 {
err := json.Unmarshal(config, &r.config)
if err != nil {
cclog.ComponentError(r.name, "Error reading config:", err.Error())
return nil, err
}
}
// Create gofish client config
for i := range r.config.ClientConfigs {
clientConfig := &r.config.ClientConfigs[i]
gofishConfig := &clientConfig.gofish
if clientConfig.Hostname == nil {
err := fmt.Errorf("client config number %v requires hostname", i)
cclog.ComponentError(r.name, err)
return nil, err
}
if clientConfig.Endpoint == nil {
err := fmt.Errorf("client config number %v requires endpoint", i)
cclog.ComponentError(r.name, err)
return nil, err
}
gofishConfig.Endpoint = *clientConfig.Endpoint
if clientConfig.Username == nil {
err := fmt.Errorf("client config number %v requires username", i)
cclog.ComponentError(r.name, err)
return nil, err
}
gofishConfig.Username = *clientConfig.Username
if clientConfig.Password == nil {
err := fmt.Errorf("client config number %v requires password", i)
cclog.ComponentError(r.name, err)
return nil, err
}
gofishConfig.Password = *clientConfig.Password
gofishConfig.Insecure = true
if clientConfig.Insecure != nil {
gofishConfig.Insecure = *clientConfig.Insecure
}
}
return r, nil
}

View File

@@ -36,16 +36,26 @@ func (r *SampleReceiver) Start() {
// or use own go routine but always make sure it exits
// as soon as it gets the signal of the r.done channel
//
// r.done = make(chan bool)
// r.wg.Add(1)
// go func() {
// for {
// select {
// case <-r.done:
// r.wg.Done()
// return
// }
// }
// r.wg.Done()
// defer r.wg.Done()
//
// // Create ticker
// ticker := time.NewTicker(30 * time.Second)
// defer ticker.Stop()
//
// for {
// readMetric()
// select {
// case <-ticker.C:
// // process ticker event -> continue
// continue
// case <-r.done:
// return
// }
// }
// }()
}

View File

@@ -15,3 +15,9 @@ CONF_DIR=/etc/cc-metric-collector
CONF_FILE=/etc/cc-metric-collector/cc-metric-collector.json
RESTART_ON_UPGRADE=true
# Golang runtime debugging. (see: https://pkg.go.dev/runtime)
# GODEBUG=gctrace=1
# Golang garbage collection target percentage
# GOGC=100

View File

@@ -0,0 +1,12 @@
Package: cc-metric-collector
Version: {VERSION}
Installed-Size: {INSTALLED_SIZE}
Architecture: {ARCH}
Maintainer: thomas.gruber@fau.de
Depends: libc6 (>= 2.2.1)
Build-Depends: debhelper-compat (= 13), git, golang-go
Description: Metric collection daemon from the ClusterCockpit suite
Homepage: https://github.com/ClusterCockpit/cc-metric-collector
Source: cc-metric-collector
Rules-Requires-Root: no

View File

@@ -75,7 +75,7 @@ case "$1" in
fi
# Start Daemon
start-stop-daemon --start -b --chdir "$WORK_DIR" --user "$CC_USER" -c "$CC_USER" --pidfile "$PID_FILE" --exec $DAEMON -- $DAEMON_OPTS
start-stop-daemon --start -b --chdir "$WORK_DIR" --user "$CC_USER" -c "$CC_USER" --pidfile "$PID_FILE" --exec $DAEMON -- $CC_OPTS
return=$?
if [ $return -eq 0 ]
then

View File

@@ -8,8 +8,8 @@ Source0: %{name}-%{version}.tar.gz
BuildRequires: go-toolset
BuildRequires: systemd-rpm-macros
# for internal LIKWID installation
BuildRequires: wget perl-Data-Dumper
# for header downloads
BuildRequires: wget
Provides: %{name} = %{version}
@@ -53,7 +53,7 @@ install -Dpm 0644 scripts/%{name}.sysusers %{buildroot}%{_sysusersdir}/%{name}.c
%files
# Binary
%attr(-,clustercockpit,clustercockpit) %{_sbindir}/%{name}
# Configuration
# Config
%dir %{_sysconfdir}/%{name}
%attr(0600,clustercockpit,clustercockpit) %config(noreplace) %{_sysconfdir}/%{name}/%{name}.json
%attr(0600,clustercockpit,clustercockpit) %config(noreplace) %{_sysconfdir}/%{name}/collectors.json
@@ -61,9 +61,9 @@ install -Dpm 0644 scripts/%{name}.sysusers %{buildroot}%{_sysusersdir}/%{name}.c
%attr(0600,clustercockpit,clustercockpit) %config(noreplace) %{_sysconfdir}/%{name}/receivers.json
%attr(0600,clustercockpit,clustercockpit) %config(noreplace) %{_sysconfdir}/%{name}/router.json
# Systemd
%{_sysusersdir}/%{name}.conf
%{_unitdir}/%{name}.service
%{_sysconfdir}/default/%{name}
%{_sysusersdir}/%{name}.conf
%changelog
* Thu Mar 03 2022 Thomas Gruber - 0.3

View File

@@ -39,7 +39,7 @@ def group_to_json(groupfile):
llist = re.split("\s+", line)
calc = llist[-1]
metric = " ".join(llist[:-1])
scope = "cpu"
scope = "hwthread"
if "BOX" in calc:
scope = "socket"
if "PWR" in calc:

View File

@@ -1,6 +1,8 @@
{
"mystdout" : {
"type" : "stdout",
"meta_as_tags" : true
"mystdout": {
"type": "stdout",
"meta_as_tags": [
"unit"
]
}
}
}

View File

@@ -148,10 +148,14 @@ type GangliaMetricConfig struct {
Unit string
Group string
Value string
Name string
}
func GetCommonGangliaConfig(point lp.CCMetric) GangliaMetricConfig {
mname := GangliaMetricRename(point.Name())
if oldname, ok := point.GetMeta("oldname"); ok {
mname = GangliaMetricRename(oldname)
}
for _, group := range CommonGangliaMetrics {
for _, metric := range group.Metrics {
if metric.Name == mname {
@@ -187,6 +191,7 @@ func GetCommonGangliaConfig(point lp.CCMetric) GangliaMetricConfig {
Tmax: metric.Tmax,
Unit: metric.Unit,
Value: valueStr,
Name: GangliaMetricRename(mname),
}
}
}
@@ -198,10 +203,15 @@ func GetCommonGangliaConfig(point lp.CCMetric) GangliaMetricConfig {
Tmax: 0,
Unit: "",
Value: "",
Name: "",
}
}
func GetGangliaConfig(point lp.CCMetric) GangliaMetricConfig {
mname := GangliaMetricRename(point.Name())
if oldname, ok := point.GetMeta("oldname"); ok {
mname = GangliaMetricRename(oldname)
}
group := ""
if g, ok := point.GetMeta("group"); ok {
group = g
@@ -254,5 +264,6 @@ func GetGangliaConfig(point lp.CCMetric) GangliaMetricConfig {
Tmax: DEFAULT_GANGLIA_METRIC_TMAX,
Unit: unit,
Value: valueStr,
Name: GangliaMetricRename(mname),
}
}

View File

@@ -39,16 +39,13 @@ func (s *GangliaSink) Write(point lp.CCMetric) error {
//var tagsstr []string
var argstr []string
// Get metric name
metricname := GangliaMetricRename(point.Name())
// Get metric config (type, value, ... in suitable format)
conf := GetCommonGangliaConfig(point)
if len(conf.Type) == 0 {
conf = GetGangliaConfig(point)
}
if len(conf.Type) == 0 {
return fmt.Errorf("metric %s has no 'value' field", metricname)
return fmt.Errorf("metric %q (Ganglia name %q) has no 'value' field", point.Name(), conf.Name)
}
if s.config.AddGangliaGroup {
@@ -70,7 +67,7 @@ func (s *GangliaSink) Write(point lp.CCMetric) error {
if s.config.AddTypeToName {
argstr = append(argstr, fmt.Sprintf("--name=%s", GangliaMetricName(point)))
} else {
argstr = append(argstr, fmt.Sprintf("--name=%s", metricname))
argstr = append(argstr, fmt.Sprintf("--name=%s", conf.Name))
}
argstr = append(argstr, fmt.Sprintf("--slope=%s", conf.Slope))
argstr = append(argstr, fmt.Sprintf("--value=%s", conf.Value))

View File

@@ -19,9 +19,9 @@ type HttpSinkConfig struct {
URL string `json:"url,omitempty"`
JWT string `json:"jwt,omitempty"`
Timeout string `json:"timeout,omitempty"`
MaxIdleConns int `json:"max_idle_connections,omitempty"`
IdleConnTimeout string `json:"idle_connection_timeout,omitempty"`
FlushDelay string `json:"flush_delay,omitempty"`
MaxRetries int `json:"max_retries,omitempty"`
}
type HttpSink struct {
@@ -32,80 +32,87 @@ type HttpSink struct {
buffer *bytes.Buffer
flushTimer *time.Timer
config HttpSinkConfig
maxIdleConns int
idleConnTimeout time.Duration
timeout time.Duration
flushDelay time.Duration
}
func (s *HttpSink) Write(m lp.CCMetric) error {
if s.buffer.Len() == 0 && s.flushDelay != 0 {
// This is the first write since the last flush, start the flushTimer!
if s.flushTimer != nil && s.flushTimer.Stop() {
cclog.ComponentDebug("HttpSink", "unexpected: the flushTimer was already running?")
}
// Run a batched flush for all lines that have arrived in the last second
s.flushTimer = time.AfterFunc(s.flushDelay, func() {
if err := s.Flush(); err != nil {
cclog.ComponentError("HttpSink", "flush failed:", err.Error())
}
})
}
p := m.ToPoint(s.config.MetaAsTags)
p := m.ToPoint(s.meta_as_tags)
s.lock.Lock()
firstWriteOfBatch := s.buffer.Len() == 0
_, err := s.encoder.Encode(p)
s.lock.Unlock() // defer does not work here as Flush() takes the lock as well
s.lock.Unlock()
if err != nil {
cclog.ComponentError(s.name, "encoding failed:", err.Error())
return err
}
// Flush synchronously if "flush_delay" is zero
if s.flushDelay == 0 {
return s.Flush()
}
return err
if firstWriteOfBatch {
if s.flushTimer == nil {
s.flushTimer = time.AfterFunc(s.flushDelay, func() {
if err := s.Flush(); err != nil {
cclog.ComponentError(s.name, "flush failed:", err.Error())
}
})
} else {
s.flushTimer.Reset(s.flushDelay)
}
}
return nil
}
func (s *HttpSink) Flush() error {
// buffer is read by client.Do, prevent concurrent modifications
// Own lock for as short as possible: the time it takes to copy the buffer.
s.lock.Lock()
defer s.lock.Unlock()
// Do not flush empty buffer
if s.buffer.Len() == 0 {
buf := make([]byte, s.buffer.Len())
copy(buf, s.buffer.Bytes())
s.buffer.Reset()
s.lock.Unlock()
if len(buf) == 0 {
return nil
}
// Create new request to send buffer
req, err := http.NewRequest(http.MethodPost, s.config.URL, s.buffer)
if err != nil {
return err
var res *http.Response
for i := 0; i < s.config.MaxRetries; i++ {
// Create new request to send buffer
req, err := http.NewRequest(http.MethodPost, s.config.URL, bytes.NewReader(buf))
if err != nil {
cclog.ComponentError(s.name, "failed to create request:", err.Error())
return err
}
// Set authorization header
if len(s.config.JWT) != 0 {
req.Header.Set("Authorization", fmt.Sprintf("Bearer %s", s.config.JWT))
}
// Do request
res, err = s.client.Do(req)
if err != nil {
cclog.ComponentError(s.name, "transport/tcp error:", err.Error())
// Wait between retries
time.Sleep(time.Duration(i+1) * (time.Second / 2))
continue
}
break
}
// Set authorization header
if len(s.config.JWT) != 0 {
req.Header.Set("Authorization", fmt.Sprintf("Bearer %s", s.config.JWT))
}
// Send
res, err := s.client.Do(req)
// Clear buffer
s.buffer.Reset()
// Handle transport/tcp errors
if err != nil {
return err
if res == nil {
return errors.New("flush failed due to repeated errors")
}
// Handle application errors
if res.StatusCode != http.StatusOK {
return errors.New(res.Status)
err := errors.New(res.Status)
cclog.ComponentError(s.name, "application error:", err.Error())
return err
}
return nil
@@ -114,7 +121,7 @@ func (s *HttpSink) Flush() error {
func (s *HttpSink) Close() {
s.flushTimer.Stop()
if err := s.Flush(); err != nil {
cclog.ComponentError("HttpSink", "flush failed:", err.Error())
cclog.ComponentError(s.name, "flush failed:", err.Error())
}
s.client.CloseIdleConnections()
}
@@ -123,10 +130,10 @@ func NewHttpSink(name string, config json.RawMessage) (Sink, error) {
s := new(HttpSink)
// Set default values
s.name = fmt.Sprintf("HttpSink(%s)", name)
s.config.MaxIdleConns = 10
s.config.IdleConnTimeout = "5s"
s.config.IdleConnTimeout = "120s" // should be larger than the measurement interval.
s.config.Timeout = "5s"
s.config.FlushDelay = "1s"
s.config.FlushDelay = "5s"
s.config.MaxRetries = 3
// Read config
if len(config) > 0 {
@@ -138,9 +145,6 @@ func NewHttpSink(name string, config json.RawMessage) (Sink, error) {
if len(s.config.URL) == 0 {
return nil, errors.New("`url` config option is required for HTTP sink")
}
if s.config.MaxIdleConns > 0 {
s.maxIdleConns = s.config.MaxIdleConns
}
if len(s.config.IdleConnTimeout) > 0 {
t, err := time.ParseDuration(s.config.IdleConnTimeout)
if err == nil {
@@ -159,8 +163,13 @@ func NewHttpSink(name string, config json.RawMessage) (Sink, error) {
s.flushDelay = t
}
}
// Create lookup map to use meta infos as tags in the output metric
s.meta_as_tags = make(map[string]bool)
for _, k := range s.config.MetaAsTags {
s.meta_as_tags[k] = true
}
tr := &http.Transport{
MaxIdleConns: s.maxIdleConns,
MaxIdleConns: 1, // We will only ever talk to one host.
IdleConnTimeout: s.idleConnTimeout,
}
s.client = &http.Client{Transport: tr, Timeout: s.timeout}

View File

@@ -6,11 +6,14 @@ import (
"encoding/json"
"errors"
"fmt"
"strings"
"time"
cclog "github.com/ClusterCockpit/cc-metric-collector/internal/ccLogger"
lp "github.com/ClusterCockpit/cc-metric-collector/internal/ccMetric"
influxdb2 "github.com/influxdata/influxdb-client-go/v2"
influxdb2Api "github.com/influxdata/influxdb-client-go/v2/api"
influxdb2ApiHttp "github.com/influxdata/influxdb-client-go/v2/api/http"
)
type InfluxAsyncSinkConfig struct {
@@ -22,19 +25,28 @@ type InfluxAsyncSinkConfig struct {
Password string `json:"password,omitempty"`
Organization string `json:"organization,omitempty"`
SSL bool `json:"ssl,omitempty"`
RetentionPol string `json:"retention_policy,omitempty"`
// Maximum number of points sent to server in single request. Default 5000
BatchSize uint `json:"batch_size,omitempty"`
// Interval, in ms, in which is buffer flushed if it has not been already written (by reaching batch size) . Default 1000ms
FlushInterval uint `json:"flush_interval,omitempty"`
FlushInterval uint `json:"flush_interval,omitempty"`
InfluxRetryInterval string `json:"retry_interval,omitempty"`
InfluxExponentialBase uint `json:"retry_exponential_base,omitempty"`
InfluxMaxRetries uint `json:"max_retries,omitempty"`
InfluxMaxRetryTime string `json:"max_retry_time,omitempty"`
CustomFlushInterval string `json:"custom_flush_interval,omitempty"`
MaxRetryAttempts uint `json:"max_retry_attempts,omitempty"`
}
type InfluxAsyncSink struct {
sink
client influxdb2.Client
writeApi influxdb2Api.WriteAPI
errors <-chan error
config InfluxAsyncSinkConfig
client influxdb2.Client
writeApi influxdb2Api.WriteAPI
errors <-chan error
config InfluxAsyncSinkConfig
influxRetryInterval uint
influxMaxRetryTime uint
customFlushInterval time.Duration
flushTimer *time.Timer
}
func (s *InfluxAsyncSink) connect() error {
@@ -53,16 +65,35 @@ func (s *InfluxAsyncSink) connect() error {
cclog.ComponentDebug(s.name, "Using URI", uri, "Org", s.config.Organization, "Bucket", s.config.Database)
clientOptions := influxdb2.DefaultOptions()
if s.config.BatchSize != 0 {
cclog.ComponentDebug(s.name, "Batch size", s.config.BatchSize)
clientOptions.SetBatchSize(s.config.BatchSize)
}
if s.config.FlushInterval != 0 {
cclog.ComponentDebug(s.name, "Flush interval", s.config.FlushInterval)
clientOptions.SetFlushInterval(s.config.FlushInterval)
}
if s.influxRetryInterval != 0 {
cclog.ComponentDebug(s.name, "MaxRetryInterval", s.influxRetryInterval)
clientOptions.SetMaxRetryInterval(s.influxRetryInterval)
}
if s.influxMaxRetryTime != 0 {
cclog.ComponentDebug(s.name, "MaxRetryTime", s.influxMaxRetryTime)
clientOptions.SetMaxRetryTime(s.influxMaxRetryTime)
}
if s.config.InfluxExponentialBase != 0 {
cclog.ComponentDebug(s.name, "Exponential Base", s.config.InfluxExponentialBase)
clientOptions.SetExponentialBase(s.config.InfluxExponentialBase)
}
if s.config.InfluxMaxRetries != 0 {
cclog.ComponentDebug(s.name, "Max Retries", s.config.InfluxMaxRetries)
clientOptions.SetMaxRetries(s.config.InfluxMaxRetries)
}
clientOptions.SetTLSConfig(
&tls.Config{
InsecureSkipVerify: true,
},
)
).SetPrecision(time.Second)
s.client = influxdb2.NewClientWithOptions(uri, auth, clientOptions)
s.writeApi = s.client.WriteAPI(s.config.Organization, s.config.Database)
ok, err := s.client.Ping(context.Background())
@@ -72,18 +103,35 @@ func (s *InfluxAsyncSink) connect() error {
if !ok {
return fmt.Errorf("connection to %s not healthy", uri)
}
s.writeApi.SetWriteFailedCallback(func(batch string, err influxdb2ApiHttp.Error, retryAttempts uint) bool {
mlist := strings.Split(batch, "\n")
cclog.ComponentError(s.name, fmt.Sprintf("Failed to write batch with %d metrics %d times (max: %d): %s", len(mlist), retryAttempts, s.config.MaxRetryAttempts, err.Error()))
return retryAttempts <= s.config.MaxRetryAttempts
})
return nil
}
func (s *InfluxAsyncSink) Write(m lp.CCMetric) error {
if s.customFlushInterval != 0 && s.flushTimer == nil {
// Run a batched flush for all lines that have arrived in the defined interval
s.flushTimer = time.AfterFunc(s.customFlushInterval, func() {
if err := s.Flush(); err != nil {
cclog.ComponentError(s.name, "flush failed:", err.Error())
}
})
}
s.writeApi.WritePoint(
m.ToPoint(s.config.MetaAsTags),
m.ToPoint(s.meta_as_tags),
)
return nil
}
func (s *InfluxAsyncSink) Flush() error {
cclog.ComponentDebug(s.name, "Flushing")
s.writeApi.Flush()
if s.customFlushInterval != 0 && s.flushTimer != nil {
s.flushTimer = nil
}
return nil
}
@@ -98,7 +146,38 @@ func NewInfluxAsyncSink(name string, config json.RawMessage) (Sink, error) {
s.name = fmt.Sprintf("InfluxSink(%s)", name)
// Set default for maximum number of points sent to server in single request.
s.config.BatchSize = 100
s.config.BatchSize = 0
s.influxRetryInterval = 0
//s.config.InfluxRetryInterval = "1s"
s.influxMaxRetryTime = 0
//s.config.InfluxMaxRetryTime = "168h"
s.config.InfluxMaxRetries = 0
s.config.InfluxExponentialBase = 0
s.config.FlushInterval = 0
s.config.CustomFlushInterval = ""
s.customFlushInterval = time.Duration(0)
s.config.MaxRetryAttempts = 1
// Default retry intervals (in seconds)
// 1 2
// 2 4
// 4 8
// 8 16
// 16 32
// 32 64
// 64 128
// 128 256
// 256 512
// 512 1024
// 1024 2048
// 2048 4096
// 4096 8192
// 8192 16384
// 16384 32768
// 32768 65536
// 65536 131072
// 131072 262144
// 262144 524288
if len(config) > 0 {
err := json.Unmarshal(config, &s.config)
@@ -106,12 +185,41 @@ func NewInfluxAsyncSink(name string, config json.RawMessage) (Sink, error) {
return nil, err
}
}
if len(s.config.Host) == 0 ||
len(s.config.Port) == 0 ||
len(s.config.Database) == 0 ||
len(s.config.Organization) == 0 ||
len(s.config.Password) == 0 {
return nil, errors.New("not all configuration variables set required by InfluxAsyncSink")
if len(s.config.Port) == 0 {
return nil, errors.New("Missing port configuration required by InfluxSink")
}
if len(s.config.Database) == 0 {
return nil, errors.New("Missing database configuration required by InfluxSink")
}
if len(s.config.Organization) == 0 {
return nil, errors.New("Missing organization configuration required by InfluxSink")
}
if len(s.config.Password) == 0 {
return nil, errors.New("Missing password configuration required by InfluxSink")
}
// Create lookup map to use meta infos as tags in the output metric
s.meta_as_tags = make(map[string]bool)
for _, k := range s.config.MetaAsTags {
s.meta_as_tags[k] = true
}
toUint := func(duration string, def uint) uint {
t, err := time.ParseDuration(duration)
if err == nil {
return uint(t.Milliseconds())
}
return def
}
s.influxRetryInterval = toUint(s.config.InfluxRetryInterval, s.influxRetryInterval)
s.influxMaxRetryTime = toUint(s.config.InfluxMaxRetryTime, s.influxMaxRetryTime)
// Use a own timer for calling Flush()
if len(s.config.CustomFlushInterval) > 0 {
t, err := time.ParseDuration(s.config.CustomFlushInterval)
if err != nil {
return nil, fmt.Errorf("invalid duration in 'custom_flush_interval': %v", err)
}
s.customFlushInterval = t
}
// Connect to InfluxDB server

View File

@@ -18,6 +18,10 @@ The `influxasync` sink uses the official [InfluxDB golang client](https://pkg.go
"organization": "myorg",
"ssl": true,
"batch_size": 200,
"retry_interval" : "1s",
"retry_exponential_base" : 2,
"max_retries": 20,
"max_retry_time" : "168h"
}
}
```
@@ -31,4 +35,10 @@ The `influxasync` sink uses the official [InfluxDB golang client](https://pkg.go
- `password`: Password for basic authentification
- `organization`: Organization in the InfluxDB
- `ssl`: Use SSL connection
- `batch_size`: batch up metrics internally, default 100
- `batch_size`: batch up metrics internally, default 100
- `retry_interval`: Base retry interval for failed write requests, default 1s
- `retry_exponential_base`: The retry interval is exponentially increased with this base, default 2
- `max_retries`: Maximal number of retry attempts
- `max_retry_time`: Maximal time to retry failed writes, default 168h (one week)
For information about the calculation of the retry interval settings, see [offical influxdb-client-go documentation](https://github.com/influxdata/influxdb-client-go#handling-of-failed-async-writes)

View File

@@ -6,54 +6,81 @@ import (
"encoding/json"
"errors"
"fmt"
"sync"
"time"
cclog "github.com/ClusterCockpit/cc-metric-collector/internal/ccLogger"
lp "github.com/ClusterCockpit/cc-metric-collector/internal/ccMetric"
influxdb2 "github.com/influxdata/influxdb-client-go/v2"
influxdb2Api "github.com/influxdata/influxdb-client-go/v2/api"
"github.com/influxdata/influxdb-client-go/v2/api/write"
)
type InfluxSinkConfig struct {
defaultSinkConfig
Host string `json:"host,omitempty"`
Port string `json:"port,omitempty"`
Database string `json:"database,omitempty"`
User string `json:"user,omitempty"`
Password string `json:"password,omitempty"`
Organization string `json:"organization,omitempty"`
SSL bool `json:"ssl,omitempty"`
RetentionPol string `json:"retention_policy,omitempty"`
}
type InfluxSink struct {
sink
client influxdb2.Client
writeApi influxdb2Api.WriteAPIBlocking
config InfluxSinkConfig
config struct {
defaultSinkConfig
Host string `json:"host,omitempty"`
Port string `json:"port,omitempty"`
Database string `json:"database,omitempty"`
User string `json:"user,omitempty"`
Password string `json:"password,omitempty"`
Organization string `json:"organization,omitempty"`
SSL bool `json:"ssl,omitempty"`
// Maximum number of points sent to server in single request. Default 100
BatchSize int `json:"batch_size,omitempty"`
// Interval, in which is buffer flushed if it has not been already written (by reaching batch size). Default 1s
FlushInterval string `json:"flush_delay,omitempty"`
}
batch []*write.Point
flushTimer *time.Timer
flushDelay time.Duration
lock sync.Mutex // Flush() runs in another goroutine, so this lock has to protect the buffer
}
// connect connects to the InfluxDB server
func (s *InfluxSink) connect() error {
var auth string
// URI options:
// * http://host:port
// * https://host:port
var uri string
if s.config.SSL {
uri = fmt.Sprintf("https://%s:%s", s.config.Host, s.config.Port)
} else {
uri = fmt.Sprintf("http://%s:%s", s.config.Host, s.config.Port)
}
// Authentication options:
// * token
// * username:password
var auth string
if len(s.config.User) == 0 {
auth = s.config.Password
} else {
auth = fmt.Sprintf("%s:%s", s.config.User, s.config.Password)
}
cclog.ComponentDebug(s.name, "Using URI", uri, "Org", s.config.Organization, "Bucket", s.config.Database)
// Set influxDB client options
clientOptions := influxdb2.DefaultOptions()
// Do not check InfluxDB certificate
clientOptions.SetTLSConfig(
&tls.Config{
InsecureSkipVerify: true,
},
)
clientOptions.SetPrecision(time.Second)
// Create new writeAPI
s.client = influxdb2.NewClientWithOptions(uri, auth, clientOptions)
s.writeApi = s.client.WriteAPIBlocking(s.config.Organization, s.config.Database)
// Check InfluxDB server accessibility
ok, err := s.client.Ping(context.Background())
if err != nil {
return err
@@ -65,39 +92,126 @@ func (s *InfluxSink) connect() error {
}
func (s *InfluxSink) Write(m lp.CCMetric) error {
err :=
s.writeApi.WritePoint(
context.Background(),
m.ToPoint(s.config.MetaAsTags),
)
return err
if len(s.batch) == 0 && s.flushDelay != 0 {
// This is the first write since the last flush, start the flushTimer!
if s.flushTimer != nil && s.flushTimer.Stop() {
cclog.ComponentDebug(s.name, "unexpected: the flushTimer was already running?")
}
// Run a batched flush for all lines that have arrived in the last flush delay interval
s.flushTimer = time.AfterFunc(
s.flushDelay,
func() {
if err := s.Flush(); err != nil {
cclog.ComponentError(s.name, "flush failed:", err.Error())
}
})
}
// Append metric to batch slice
p := m.ToPoint(s.meta_as_tags)
s.lock.Lock()
s.batch = append(s.batch, p)
s.lock.Unlock()
// Flush synchronously if "flush_delay" is zero
if s.flushDelay == 0 {
return s.Flush()
}
// Flush if batch size is reached
if len(s.batch) == s.config.BatchSize {
return s.Flush()
}
return nil
}
// Flush sends all metrics buffered in batch slice to InfluxDB server
func (s *InfluxSink) Flush() error {
// Lock access to batch slice
s.lock.Lock()
defer s.lock.Unlock()
// Nothing to do, batch slice is empty
if len(s.batch) == 0 {
return nil
}
// Send metrics from batch slice
err := s.writeApi.WritePoint(context.Background(), s.batch...)
if err != nil {
cclog.ComponentError(s.name, "flush failed:", err.Error())
return err
}
// Clear batch slice
for i := range s.batch {
s.batch[i] = nil
}
s.batch = s.batch[:0]
return nil
}
func (s *InfluxSink) Close() {
cclog.ComponentDebug(s.name, "Closing InfluxDB connection")
s.flushTimer.Stop()
s.Flush()
s.client.Close()
}
// NewInfluxSink create a new InfluxDB sink
func NewInfluxSink(name string, config json.RawMessage) (Sink, error) {
s := new(InfluxSink)
s.name = fmt.Sprintf("InfluxSink(%s)", name)
// Set config default values
s.config.BatchSize = 100
s.config.FlushInterval = "1s"
// Read config
if len(config) > 0 {
err := json.Unmarshal(config, &s.config)
if err != nil {
return nil, err
}
}
if len(s.config.Host) == 0 ||
len(s.config.Port) == 0 ||
len(s.config.Database) == 0 ||
len(s.config.Organization) == 0 ||
len(s.config.Password) == 0 {
return nil, errors.New("not all configuration variables set required by InfluxSink")
if len(s.config.Host) == 0 {
return nil, errors.New("Missing host configuration required by InfluxSink")
}
if len(s.config.Port) == 0 {
return nil, errors.New("Missing port configuration required by InfluxSink")
}
if len(s.config.Database) == 0 {
return nil, errors.New("Missing database configuration required by InfluxSink")
}
if len(s.config.Organization) == 0 {
return nil, errors.New("Missing organization configuration required by InfluxSink")
}
if len(s.config.Password) == 0 {
return nil, errors.New("Missing password configuration required by InfluxSink")
}
// Create lookup map to use meta infos as tags in the output metric
s.meta_as_tags = make(map[string]bool)
for _, k := range s.config.MetaAsTags {
s.meta_as_tags[k] = true
}
// Configure flush delay duration
if len(s.config.FlushInterval) > 0 {
t, err := time.ParseDuration(s.config.FlushInterval)
if err == nil {
s.flushDelay = t
}
}
// allocate batch slice
s.batch = make([]*write.Point, 0, s.config.BatchSize)
// Connect to InfluxDB server
if err := s.connect(); err != nil {

View File

@@ -17,6 +17,8 @@ The `influxdb` sink uses the official [InfluxDB golang client](https://pkg.go.de
"password" : "examplepw",
"organization": "myorg",
"ssl": true,
"flush_delay" : "1s",
"batch_size" : 100
}
}
```
@@ -29,4 +31,7 @@ The `influxdb` sink uses the official [InfluxDB golang client](https://pkg.go.de
- `user`: Username for basic authentification
- `password`: Password for basic authentification
- `organization`: Organization in the InfluxDB
- `ssl`: Use SSL connection
- `ssl`: Use SSL connection
- `flush_delay`: Group metrics coming in to a single batch
- `batch_size`: Maximal batch size

View File

@@ -124,24 +124,21 @@ func (s *LibgangliaSink) Write(point lp.CCMetric) error {
return s.cstrCache[key]
}
// Get metric name
metricname := GangliaMetricRename(point.Name())
conf := GetCommonGangliaConfig(point)
if len(conf.Type) == 0 {
conf = GetGangliaConfig(point)
}
if len(conf.Type) == 0 {
return fmt.Errorf("metric %s has no 'value' field", metricname)
return fmt.Errorf("metric %q (Ganglia name %q) has no 'value' field", point.Name(), conf.Name)
}
if s.config.AddTypeToName {
metricname = GangliaMetricName(point)
conf.Name = GangliaMetricName(point)
}
c_value = C.CString(conf.Value)
c_type = lookup(conf.Type)
c_name = lookup(metricname)
c_name = lookup(conf.Name)
// Add unit
unit := ""

View File

@@ -5,13 +5,13 @@ import (
)
type defaultSinkConfig struct {
MetaAsTags bool `json:"meta_as_tags,omitempty"`
Type string `json:"type"`
MetaAsTags []string `json:"meta_as_tags,omitempty"`
Type string `json:"type"`
}
type sink struct {
meta_as_tags bool // Use meta data tags as tags
name string // Name of the sink
meta_as_tags map[string]bool // Use meta data tags as tags
name string // Name of the sink
}
type Sink interface {

View File

@@ -5,6 +5,7 @@ import (
"encoding/json"
"errors"
"fmt"
"sync"
"time"
cclog "github.com/ClusterCockpit/cc-metric-collector/internal/ccLogger"
@@ -15,11 +16,12 @@ import (
type NatsSinkConfig struct {
defaultSinkConfig
Host string `json:"host,omitempty"`
Port string `json:"port,omitempty"`
Database string `json:"database,omitempty"`
User string `json:"user,omitempty"`
Password string `json:"password,omitempty"`
Host string `json:"host,omitempty"`
Port string `json:"port,omitempty"`
Subject string `json:"subject,omitempty"`
User string `json:"user,omitempty"`
Password string `json:"password,omitempty"`
FlushDelay string `json:"flush_delay,omitempty"`
}
type NatsSink struct {
@@ -28,6 +30,10 @@ type NatsSink struct {
encoder *influx.Encoder
buffer *bytes.Buffer
config NatsSinkConfig
lock sync.Mutex
flushDelay time.Duration
flushTimer *time.Timer
}
func (s *NatsSink) connect() error {
@@ -54,37 +60,53 @@ func (s *NatsSink) connect() error {
}
func (s *NatsSink) Write(m lp.CCMetric) error {
if s.client != nil {
_, err := s.encoder.Encode(m.ToPoint(s.config.MetaAsTags))
if err != nil {
cclog.ComponentError(s.name, "Write:", err.Error())
return err
}
s.lock.Lock()
_, err := s.encoder.Encode(m.ToPoint(s.meta_as_tags))
s.lock.Unlock()
if err != nil {
cclog.ComponentError(s.name, "Write:", err.Error())
return err
}
if s.flushDelay == 0 {
s.Flush()
} else if s.flushTimer == nil {
s.flushTimer = time.AfterFunc(s.flushDelay, func() {
s.Flush()
})
} else {
s.flushTimer.Reset(s.flushDelay)
}
return nil
}
func (s *NatsSink) Flush() error {
if s.client != nil {
if err := s.client.Publish(s.config.Database, s.buffer.Bytes()); err != nil {
cclog.ComponentError(s.name, "Flush:", err.Error())
return err
}
s.buffer.Reset()
s.lock.Lock()
buf := append([]byte{}, s.buffer.Bytes()...) // copy bytes
s.buffer.Reset()
s.lock.Unlock()
if len(buf) == 0 {
return nil
}
if err := s.client.Publish(s.config.Subject, buf); err != nil {
cclog.ComponentError(s.name, "Flush:", err.Error())
return err
}
return nil
}
func (s *NatsSink) Close() {
if s.client != nil {
cclog.ComponentDebug(s.name, "Close")
s.client.Close()
}
cclog.ComponentDebug(s.name, "Close")
s.client.Close()
}
func NewNatsSink(name string, config json.RawMessage) (Sink, error) {
s := new(NatsSink)
s.name = fmt.Sprintf("NatsSink(%s)", name)
s.flushDelay = 10 * time.Second
if len(config) > 0 {
err := json.Unmarshal(config, &s.config)
if err != nil {
@@ -94,9 +116,14 @@ func NewNatsSink(name string, config json.RawMessage) (Sink, error) {
}
if len(s.config.Host) == 0 ||
len(s.config.Port) == 0 ||
len(s.config.Database) == 0 {
len(s.config.Subject) == 0 {
return nil, errors.New("not all configuration variables set required by NatsSink")
}
// Create lookup map to use meta infos as tags in the output metric
s.meta_as_tags = make(map[string]bool)
for _, k := range s.config.MetaAsTags {
s.meta_as_tags[k] = true
}
// Setup Influx line protocol
s.buffer = &bytes.Buffer{}
s.buffer.Grow(1025)
@@ -105,7 +132,17 @@ func NewNatsSink(name string, config json.RawMessage) (Sink, error) {
s.encoder.SetMaxLineBytes(1024)
// Setup infos for connection
if err := s.connect(); err != nil {
return nil, fmt.Errorf("Unable to connect: %v", err)
return nil, fmt.Errorf("unable to connect: %v", err)
}
s.flushTimer = nil
if len(s.config.FlushDelay) != 0 {
var err error
s.flushDelay, err = time.ParseDuration(s.config.FlushDelay)
if err != nil {
return nil, err
}
}
return s, nil
}

View File

@@ -10,14 +10,14 @@ import (
)
type SampleSinkConfig struct {
// defines JSON tags for 'type' and 'meta_as_tags'
// defines JSON tags for 'type' and 'meta_as_tags' (string list)
// See: metricSink.go
defaultSinkConfig
// Additional config options, for SampleSink
}
type SampleSink struct {
// declares elements 'name' and 'meta_as_tags'
// declares elements 'name' and 'meta_as_tags' (string to bool map!)
sink
config SampleSinkConfig // entry point to the SampleSinkConfig
}
@@ -28,6 +28,7 @@ type SampleSink struct {
// Code to submit a single CCMetric to the sink
func (s *SampleSink) Write(point lp.CCMetric) error {
// based on s.meta_as_tags use meta infos as tags
log.Print(point)
return nil
}
@@ -62,6 +63,12 @@ func NewSampleSink(name string, config json.RawMessage) (Sink, error) {
}
}
// Create lookup map to use meta infos as tags in the output metric
s.meta_as_tags = make(map[string]bool)
for _, k := range s.config.MetaAsTags {
s.meta_as_tags[k] = true
}
// Check if all required fields in the config are set
// E.g. use 'len(s.config.Option) > 0' for string settings

View File

@@ -63,7 +63,11 @@ func NewStdoutSink(name string, config json.RawMessage) (Sink, error) {
s.output = f
}
}
s.meta_as_tags = s.config.MetaAsTags
// Create lookup map to use meta infos as tags in the output metric
s.meta_as_tags = make(map[string]bool)
for _, k := range s.config.MetaAsTags {
s.meta_as_tags[k] = true
}
return s, nil
}