ClusterCockpit/cc-metric-collector
1
0
Fork 0
mirror of https://github.com/ClusterCockpit/cc-metric-collector.git synced 2025-07-31 08:56:06 +02:00
Code Issues Packages Projects Releases Wiki Activity

Merge develop into main (#109)

Browse Source 
* Add cpu_used (all-cpu_idle) to CpustatCollector

* Update to line-protocol/v2

* Update runonce.yml with Golang 1.20

* Update fsnotify in LIKWID Collector

* Use not a pointer to line-protocol.Encoder

* Simplify Makefile

* Use only as many arguments as required

* Allow sum function to handle non float types

* Allow values to be a slice of type float64, float32, int, int64, int32, bool

* Use generic function to simplify code

* Add missing case for type []int32

* Use generic function to compute minimum

* Use generic function to compute maximum

* Use generic function to compute average

* Add error value to sumAnyType

* Use generic function to compute median

* For older versions of go slices is not part of the installation

* Remove old entries from go.sum

* Use simpler sort function

* Compute metrics ib_total and ib_total_pkts

* Add aggregated metrics.
Add missing units

* Update likwidMetric.go

Fixes a potential bug when `fsnotify.NewWatcher()` fails with an error

* Completly avoid memory allocations in infinibandMetric read()

* Fixed initialization: Initalization and measurements should run in the same thread

* Add safe.directory to Release action

* Fix path after installation to /usr/bin after installation

* ioutil.ReadFile is deprecated: As of Go 1.16, this function simply calls os.ReadFile

* Switch to package slices from the golang 1.21 default library

* Read file line by line

* Read file line by line

* Read file line by line

* Use CamelCase

* Use CamelCase

* Fix function getNumaDomain, it always returned 0

* Avoid type conversion by using Atoi
Avoid copying structs by using pointer access
Increase readability with CamelCase variable names

* Add caching

* Cache CpuData

* Cleanup

* Use init function to initalize cache structure to avoid multi threading problems

* Reuse information from /proc/cpuinfo

* Avoid slice cloning. Directly use the cache

* Add DieList

* Add NumaDomainList and SMTList

* Cleanup

* Add comment

* Lookup core ID from /sys/devices/system/cpu, /proc/cpuinfo is not portable

* Lookup all information from /sys/devices/system/cpu, /proc/cpuinfo is not portable

* Correctly handle lists from /sys

* Add Simultaneous Multithreading siblings

* Replace deprecated thread_siblings_list by core_cpus_list

* Reduce number of required slices

* Allow to send total values per core, socket and node

* Send all metrics with same time stamp
calcEventsetMetrics does only computiation, counter measurement is done before

* Input parameters should be float64 when evaluating to float64

* Send all metrics with same time stamp
calcGlobalMetrics does only computiation, counter measurement is done before

* Remove unused variable gmresults

* Add comments

* Updated go packages

* Add build with golang 1.21

* Switch to checkout action version 4

* Switch to setup-go action version 4

* Add workflow_dispatch to allow manual run of workflow

* Add workflow_dispatch to allow manual run of workflow

* Add release build jobs to runonce.yml

* Switch to golang 1.20 for RHEL based distributions

* Use dnf to download golang

* Remove golang versions before 1.20

* Upgrade Ubuntu focal -> jammy

* Pipe golang tar package directly to tar

* Update golang version

* Fix Ubuntu version number

* Add links to ipmi and redfish receivers

* Fix http server addr format

* github.com/influxdata/line-protocol -> github.com/influxdata/line-protocol/v2/lineprotocol

* Corrected spelling

* Add some comments

* github.com/influxdata/line-protocol -> github.com/influxdata/line-protocol/v2/lineprotocol

* Allow other fields not only field "value"

* Add some basic debugging documentation

* Add some basic debugging documentation

* Use a lock for the flush timer

* Add tags in lexical order as required by AddTag()

* Only access meta data, when it gets used as tag

* Use slice to store lexialicly orderd key value pairs

* Increase golang version requirement to 1.20.

* Avoid package cmp to allow builds with golang v1.20

* Fix: Error NVML library not found did crash
cc-metric-collector with "SIGSEGV: segmentation violation"

* Add config option idle_timeout

* Add basic authentication support

* Add basic authentication support

* Avoid unneccessary memory allocations

* Add documentation for send_*_total values

* Use generic package maps to clone maps

* Reuse flush timer

* Add Influx client options

* Reuse ccTopology functionality

* Do not store unused topology information

* Add batch_size config

* Cleanup

* Use stype and stype-id for the NIC in NetstatCollector

* Wait for concurrent flush operations to finish

* Be more verbose in error messages

* Reverted previous changes.
Made the code to complex without much advantages

* Use line protocol encoder

* Go pkg update

* Stop flush timer, when immediatelly flushing

* Fix: Corrected unlock access to batch slice

* Add config option to specify whether to use GZip compression in influx write requests

* Add asynchron send of encoder metrics

* Use DefaultServeMux instead of github.com/gorilla/mux

* Add config option for HTTP keep-alives

* Be more strict, when parsing json

* Add config option for HTTP request timeout and Retry interval

* Allow more then one background send operation

* Fix %sysusers_create_package args (#108)

%sysusers_create_package requires two arguments. See: https://github.com/systemd/systemd/blob/main/src/rpm/macros.systemd.in#L165

* Add nfsiostat to list of collectors

---------

Co-authored-by: Holger Obermaier <40787752+ho-ob@users.noreply.github.com>
Co-authored-by: Holger Obermaier <holgerob@gmx.de>
Co-authored-by: Obihörnchen <obihoernchende@gmail.com>
This commit is contained in:
Thomas Gruber
2023-12-04 12:21:26 +01:00
committed by GitHub
parent 9df1054e32
commit 6ab45dd3ec
39 changed files with 1835 additions and 1030 deletions
Download Patch File Download Diff File Expand all files Collapse all files

130
collectors/cpufreqMetric.go
View File

@@ -11,22 +11,13 @@ import (
cclog "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
lp "github.com/ClusterCockpit/cc-metric-collector/pkg/ccMetric"
"github.com/ClusterCockpit/cc-metric-collector/pkg/ccTopology"
"golang.org/x/sys/unix"
)
type CPUFreqCollectorTopology struct {
processor string // logical processor number (continuous, starting at 0)
coreID string // socket local core ID
coreID_int int64
physicalPackageID string // socket / package ID
physicalPackageID_int int64
numPhysicalPackages string // number of sockets / packages
numPhysicalPackages_int int64
isHT bool
numNonHT string // number of non hyper-threading processors
numNonHT_int int64
scalingCurFreqFile string
tagSet map[string]string
scalingCurFreqFile string
tagSet map[string]string
}
// CPUFreqCollector
@@ -64,112 +55,38 @@ func (m *CPUFreqCollector) Init(config json.RawMessage) error {
"unit": "Hz",
}
// Loop for all CPU directories
baseDir := "/sys/devices/system/cpu"
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)
}
if cpuDirs == nil {
return fmt.Errorf("unable to find any files with pattern '%s'", globPattern)
}
m.topology = make([]CPUFreqCollectorTopology, 0)
for _, c := range ccTopology.CpuData() {
// Initialize CPU topology
m.topology = make([]CPUFreqCollectorTopology, len(cpuDirs))
for _, cpuDir := range cpuDirs {
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)
}
// Read package ID
physicalPackageIDFile := filepath.Join(cpuDir, "topology", "physical_package_id")
line, err := os.ReadFile(physicalPackageIDFile)
if err != nil {
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)
}
// Read core ID
coreIDFile := filepath.Join(cpuDir, "topology", "core_id")
line, err = os.ReadFile(coreIDFile)
if err != nil {
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)
// Skip hyper threading CPUs
if c.CpuID != c.CoreCPUsList[0] {
continue
}
// Check access to current frequency file
scalingCurFreqFile := filepath.Join(cpuDir, "cpufreq", "scaling_cur_freq")
err = unix.Access(scalingCurFreqFile, unix.R_OK)
scalingCurFreqFile := filepath.Join("/sys/devices/system/cpu", fmt.Sprintf("cpu%d", c.CpuID), "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)
}
t := &m.topology[processor_int]
t.processor = processor
t.physicalPackageID = physicalPackageID
t.physicalPackageID_int = physicalPackageID_int
t.coreID = coreID
t.coreID_int = coreID_int
t.scalingCurFreqFile = scalingCurFreqFile
}
// is processor a hyper-thread?
coreSeenBefore := make(map[string]bool)
for i := range m.topology {
t := &m.topology[i]
globalID := t.physicalPackageID + ":" + t.coreID
t.isHT = coreSeenBefore[globalID]
coreSeenBefore[globalID] = true
}
// number of non hyper-thread cores and packages / sockets
var numNonHT_int int64 = 0
PhysicalPackageIDs := make(map[int64]struct{})
for i := range m.topology {
t := &m.topology[i]
if !t.isHT {
numNonHT_int++
}
PhysicalPackageIDs[t.physicalPackageID_int] = struct{}{}
}
numPhysicalPackageID_int := int64(len(PhysicalPackageIDs))
numPhysicalPackageID := fmt.Sprint(numPhysicalPackageID_int)
numNonHT := fmt.Sprint(numNonHT_int)
for i := range m.topology {
t := &m.topology[i]
t.numPhysicalPackages = numPhysicalPackageID
t.numPhysicalPackages_int = numPhysicalPackageID_int
t.numNonHT = numNonHT
t.numNonHT_int = numNonHT_int
t.tagSet = map[string]string{
"type": "hwthread",
"type-id": t.processor,
"package_id": t.physicalPackageID,
}
m.topology = append(m.topology,
CPUFreqCollectorTopology{
tagSet: map[string]string{
"type": "hwthread",
"type-id": fmt.Sprint(c.CpuID),
"package_id": fmt.Sprint(c.Socket),
},
scalingCurFreqFile: scalingCurFreqFile,
},
)
}
// Initialized
cclog.ComponentDebug(
m.name,
"initialized",
numPhysicalPackageID_int, "physical packages,",
len(cpuDirs), "CPUs,",
numNonHT, "non-hyper-threading CPUs")
len(m.topology), "non-hyper-threading CPUs")
m.init = true
return nil
}
@@ -184,11 +101,6 @@ func (m *CPUFreqCollector) Read(interval time.Duration, output chan lp.CCMetric)
for i := range m.topology {
t := &m.topology[i]
// skip hyper-threads
if t.isHT {
continue
}
// Read current frequency
line, err := os.ReadFile(t.scalingCurFreqFile)
if err != nil {