Compare commits

..

23 Commits

Author SHA1 Message Date
Thomas Roehl
77fceb17d1 Update rapl collector with powercap limits 2025-04-04 12:04:27 +02:00
Thomas Roehl
b5520efc25 Fix artifacts in netstat collector of not done cc-lib switch 2025-03-15 04:02:26 +01:00
Thomas Roehl
d2b1bad1b8 Fix artifacts of not done cc-lib switch 2025-03-15 04:01:01 +01:00
brinkcoder
0e57c8db1c Add derived_values for numastats (#134)
* Check creation of CCMessage in NATS receiver

* add derived_values for numastats

* change to ccMessage

* remove vim command artefact

---------

Co-authored-by: Thomas Roehl <thomas.roehl@fau.de>
Co-authored-by: exterr2f <Robert.Externbrink@rub.de>
Co-authored-by: Thomas Gruber <Thomas.Roehl@googlemail.com>
2025-02-19 11:35:32 +01:00
brinkcoder
f2f38c81af Add exclude_devices to iostat (#133)
* Check creation of CCMessage in NATS receiver

* add exclude_device for iostatMetric

* add md file

---------

Co-authored-by: Thomas Roehl <thomas.roehl@fau.de>
Co-authored-by: exterr2f <Robert.Externbrink@rub.de>
Co-authored-by: Thomas Gruber <Thomas.Roehl@googlemail.com>
2025-02-19 11:34:56 +01:00
brinkcoder
f9acc51a50 Add derived values for nfsiostat (#132)
* Check creation of CCMessage in NATS receiver

* add derived_values for nfsiostatMetric

---------

Co-authored-by: Thomas Roehl <thomas.roehl@fau.de>
Co-authored-by: exterr2f <Robert.Externbrink@rub.de>
Co-authored-by: Thomas Gruber <Thomas.Roehl@googlemail.com>
2025-02-19 11:34:06 +01:00
brinkcoder
87346e2eae Fix excluded metrics for diskstat and add exclude_mounts (#131)
* Check creation of CCMessage in NATS receiver

* fix excluded metrics and add optional mountpoint exclude

---------

Co-authored-by: Thomas Roehl <thomas.roehl@fau.de>
Co-authored-by: exterr2f <Robert.Externbrink@rub.de>
Co-authored-by: Thomas Gruber <Thomas.Roehl@googlemail.com>
2025-02-19 11:33:13 +01:00
brinkcoder
0f92f10b66 Add optional interface alias in netstat (#130)
* Check creation of CCMessage in NATS receiver

* add optional interface aliases for netstatMetric

* small fix

---------

Co-authored-by: Thomas Roehl <thomas.roehl@fau.de>
Co-authored-by: exterr2f <Robert.Externbrink@rub.de>
Co-authored-by: Thomas Gruber <Thomas.Roehl@googlemail.com>
2025-02-19 11:32:15 +01:00
Michael Panzlaff
6901b06e44 Rename 'process_message' to 'process_messages' in metricRouter config
This makes the behavior more consistent with the other modules, which
have their MessageProcessor named 'process_messages'. This most likely
was just a typo.
2025-02-03 15:23:51 +01:00
Thomas Roehl
7b343d0bab Use CCMessage FromBytes instead of Influx's decoder 2024-12-27 15:22:59 +00:00
Thomas Roehl
7d3180b526 Check creation of CCMessage in NATS receiver 2024-12-27 15:00:48 +00:00
Thomas Roehl
70a6afc549 Generate HUGO inputs out of Markdown files 2024-12-23 17:55:48 +01:00
Thomas Roehl
e02a018327 Mark all JSON config fields of message processor as omitempty 2024-12-23 17:52:34 +01:00
Thomas Roehl
bcecdd033b Fix documentation of RAPL collector 2024-12-23 17:51:43 +01:00
Thomas Roehl
2645ffeff3 Merge branch 'main' into develop 2024-12-21 02:39:08 +01:00
Thomas Roehl
e968aa1991 Fix wrongly named packages 2024-12-20 20:33:10 +01:00
Thomas Gruber
d2a38e3844 Merge branch 'main' into develop 2024-12-20 20:27:48 +01:00
Thomas Roehl
1f35f6d3ca Fix wrongly named packages 2024-12-20 20:26:38 +01:00
Thomas Roehl
7e6870c7b3 Add golang-race for UBI9 and Alma9 2024-12-20 20:15:59 +01:00
Thomas Roehl
d881093524 Install go-toolkit to fulfill build requirements for RPM 2024-12-20 20:12:03 +01:00
Thomas Roehl
c01096c157 use go-toolkit for RPM builds 2024-12-20 18:49:28 +01:00
Thomas Roehl
3d70c8afc9 Remove condition around BuildRequires and use go-toolkit for RPM builds 2024-12-20 18:43:21 +01:00
Thomas Roehl
7ee85a07dc Remove go-toolkit as build requirement for RPM builds if run in CI 2024-12-20 18:28:32 +01:00
115 changed files with 8682 additions and 616 deletions

View File

@@ -1,10 +1,8 @@
{
"sinks-file": ".github/ci-sinks.json",
"collectors-file" : ".github/ci-collectors.json",
"receivers-file" : ".github/ci-receivers.json",
"router-file" : ".github/ci-router.json",
"main" : {
"interval": "5s",
"duration": "1s"
}
"sinks": ".github/ci-sinks.json",
"collectors" : ".github/ci-collectors.json",
"receivers" : ".github/ci-receivers.json",
"router" : ".github/ci-router.json",
"interval": "5s",
"duration": "1s"
}

View File

@@ -11,10 +11,10 @@ on:
jobs:
#
# Job build-latest
# Build on latest Ubuntu using latest golang version
# Job build-1-21
# Build on latest Ubuntu using golang version 1.21
#
build-latest:
build-1-21:
runs-on: ubuntu-latest
steps:
# See: https://github.com/marketplace/actions/checkout
@@ -29,7 +29,58 @@ jobs:
uses: actions/setup-go@v5
with:
go-version: '1.21'
check-latest: true
- name: Build MetricCollector
run: make
- name: Run MetricCollector once
run: ./cc-metric-collector --once --config .github/ci-config.json
#
# Job build-1-22
# Build on latest Ubuntu using golang version 1.22
#
build-1-22:
runs-on: ubuntu-latest
steps:
# See: https://github.com/marketplace/actions/checkout
# Checkout git repository and submodules
- name: Checkout
uses: actions/checkout@v4
with:
submodules: recursive
# See: https://github.com/marketplace/actions/setup-go-environment
- name: Setup Golang
uses: actions/setup-go@v5
with:
go-version: '1.22'
- name: Build MetricCollector
run: make
- name: Run MetricCollector once
run: ./cc-metric-collector --once --config .github/ci-config.json
#
# Job build-1-23
# Build on latest Ubuntu using golang version 1.23
#
build-1-23:
runs-on: ubuntu-latest
steps:
# See: https://github.com/marketplace/actions/checkout
# Checkout git repository and submodules
- name: Checkout
uses: actions/checkout@v4
with:
submodules: recursive
# See: https://github.com/marketplace/actions/setup-go-environment
- name: Setup Golang
uses: actions/setup-go@v5
with:
go-version: '1.23'
- name: Build MetricCollector
run: make

View File

@@ -1,17 +1,6 @@
<!--
---
title: cc-metric-collector
description: Metric collecting node agent
categories: [cc-metric-collector]
tags: ['Admin']
weight: 2
hugo_path: docs/reference/cc-metric-collector/_index.md
---
-->
# cc-metric-collector
A node agent for measuring, processing and forwarding node level metrics. It is part of the [ClusterCockpit ecosystem](https://clustercockpit.org/docs/overview/).
A node agent for measuring, processing and forwarding node level metrics. It is part of the [ClusterCockpit ecosystem](./docs/introduction.md).
The metric collector sends (and receives) metric in the [InfluxDB line protocol](https://docs.influxdata.com/influxdb/cloud/reference/syntax/line-protocol/) as it provides flexibility while providing a separation between tags (like index columns in relational databases) and fields (like data columns).
@@ -46,8 +35,8 @@ The `interval` defines how often the metrics should be read and send to the sink
See the component READMEs for their configuration:
* [`collectors`](./collectors/README.md)
* [`sinks`](https://github.com/ClusterCockpit/cc-lib/blob/main/sinks/README.md)
* [`receivers`](https://github.com/ClusterCockpit/cc-lib/blob/main/receivers/README.md)
* [`sinks`](./sinks/README.md)
* [`receivers`](./receivers/README.md)
* [`router`](./internal/metricRouter/README.md)
# Installation

View File

@@ -7,24 +7,39 @@ import (
"os/signal"
"syscall"
"github.com/ClusterCockpit/cc-lib/receivers"
"github.com/ClusterCockpit/cc-lib/sinks"
"github.com/ClusterCockpit/cc-metric-collector/collectors"
"github.com/ClusterCockpit/cc-metric-collector/receivers"
"github.com/ClusterCockpit/cc-metric-collector/sinks"
// "strings"
"sync"
"time"
ccconf "github.com/ClusterCockpit/cc-lib/ccConfig"
cclog "github.com/ClusterCockpit/cc-lib/ccLogger"
lp "github.com/ClusterCockpit/cc-lib/ccMessage"
mr "github.com/ClusterCockpit/cc-metric-collector/internal/metricRouter"
cclog "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
mct "github.com/ClusterCockpit/cc-metric-collector/pkg/multiChanTicker"
)
type CentralConfigFile struct {
Interval string `json:"interval"`
Duration string `json:"duration"`
Interval string `json:"interval"`
Duration string `json:"duration"`
CollectorConfigFile string `json:"collectors"`
RouterConfigFile string `json:"router"`
SinkConfigFile string `json:"sinks"`
ReceiverConfigFile string `json:"receivers,omitempty"`
}
func LoadCentralConfiguration(file string, config *CentralConfigFile) error {
configFile, err := os.Open(file)
if err != nil {
cclog.Error(err.Error())
return err
}
defer configFile.Close()
jsonParser := json.NewDecoder(configFile)
err = jsonParser.Decode(config)
return err
}
type RuntimeConfig struct {
@@ -72,7 +87,7 @@ func ReadCli() map[string]string {
cfg := flag.String("config", "./config.json", "Path to configuration file")
logfile := flag.String("log", "stderr", "Path for logfile")
once := flag.Bool("once", false, "Run all collectors only once")
loglevel := flag.String("loglevel", "info", "Set log level")
debug := flag.Bool("debug", false, "Activate debug output")
flag.Parse()
m = make(map[string]string)
m["configfile"] = *cfg
@@ -82,7 +97,12 @@ func ReadCli() map[string]string {
} else {
m["once"] = "false"
}
m["loglevel"] = *loglevel
if *debug {
m["debug"] = "true"
cclog.SetDebug()
} else {
m["debug"] = "false"
}
return m
}
@@ -147,15 +167,8 @@ func mainFunc() int {
CliArgs: ReadCli(),
}
// Set loglevel based on command line input.
cclog.Init(rcfg.CliArgs["loglevel"], false)
// Init ccConfig with configuration file
ccconf.Init(rcfg.CliArgs["configfile"])
// Load and check configuration
main := ccconf.GetPackageConfig("main")
err = json.Unmarshal(main, &rcfg.ConfigFile)
err = LoadCentralConfiguration(rcfg.CliArgs["configfile"], &rcfg.ConfigFile)
if err != nil {
cclog.Error("Error reading configuration file ", rcfg.CliArgs["configfile"], ": ", err.Error())
return 1
@@ -191,41 +204,38 @@ func mainFunc() int {
return 1
}
routerConf := ccconf.GetPackageConfig("router")
if len(routerConf) == 0 {
if len(rcfg.ConfigFile.RouterConfigFile) == 0 {
cclog.Error("Metric router configuration file must be set")
return 1
}
sinkConf := ccconf.GetPackageConfig("sinks")
if len(sinkConf) == 0 {
if len(rcfg.ConfigFile.SinkConfigFile) == 0 {
cclog.Error("Sink configuration file must be set")
return 1
}
collectorConf := ccconf.GetPackageConfig("collectors")
if len(collectorConf) == 0 {
if len(rcfg.ConfigFile.CollectorConfigFile) == 0 {
cclog.Error("Metric collector configuration file must be set")
return 1
}
// Set log file
// if logfile := rcfg.CliArgs["logfile"]; logfile != "stderr" {
// cclog.SetOutput(logfile)
// }
if logfile := rcfg.CliArgs["logfile"]; logfile != "stderr" {
cclog.SetOutput(logfile)
}
// Creat new multi channel ticker
rcfg.MultiChanTicker = mct.NewTicker(rcfg.Interval)
// Create new metric router
rcfg.MetricRouter, err = mr.New(rcfg.MultiChanTicker, &rcfg.Sync, routerConf)
rcfg.MetricRouter, err = mr.New(rcfg.MultiChanTicker, &rcfg.Sync, rcfg.ConfigFile.RouterConfigFile)
if err != nil {
cclog.Error(err.Error())
return 1
}
// Create new sink
rcfg.SinkManager, err = sinks.New(&rcfg.Sync, sinkConf)
rcfg.SinkManager, err = sinks.New(&rcfg.Sync, rcfg.ConfigFile.SinkConfigFile)
if err != nil {
cclog.Error(err.Error())
return 1
@@ -237,7 +247,7 @@ func mainFunc() int {
rcfg.MetricRouter.AddOutput(RouterToSinksChannel)
// Create new collector manager
rcfg.CollectManager, err = collectors.New(rcfg.MultiChanTicker, rcfg.Duration, &rcfg.Sync, collectorConf)
rcfg.CollectManager, err = collectors.New(rcfg.MultiChanTicker, rcfg.Duration, &rcfg.Sync, rcfg.ConfigFile.CollectorConfigFile)
if err != nil {
cclog.Error(err.Error())
return 1
@@ -249,9 +259,8 @@ func mainFunc() int {
rcfg.MetricRouter.AddCollectorInput(CollectToRouterChannel)
// Create new receive manager
receiveConf := ccconf.GetPackageConfig("receivers")
if len(receiveConf) > 0 {
rcfg.ReceiveManager, err = receivers.New(&rcfg.Sync, receiveConf)
if len(rcfg.ConfigFile.ReceiverConfigFile) > 0 {
rcfg.ReceiveManager, err = receivers.New(&rcfg.Sync, rcfg.ConfigFile.ReceiverConfigFile)
if err != nil {
cclog.Error(err.Error())
return 1

View File

@@ -1,14 +1,3 @@
<!--
---
title: Metric Collectors
description: Metric collectors for cc-metric-collector
categories: [cc-metric-collector]
tags: ['Admin']
weight: 2
hugo_path: docs/reference/cc-metric-collector/collectors/_index.md
---
-->
# CCMetric collectors
This folder contains the collectors for the cc-metric-collector.
@@ -34,6 +23,7 @@ In contrast to the configuration files for sinks and receivers, the collectors c
* [`loadavg`](./loadavgMetric.md)
* [`netstat`](./netstatMetric.md)
* [`ibstat`](./infinibandMetric.md)
* [`ibstat_perfquery`](./infinibandPerfQueryMetric.md)
* [`tempstat`](./tempMetric.md)
* [`lustrestat`](./lustreMetric.md)
* [`likwid`](./likwidMetric.md)
@@ -43,10 +33,8 @@ In contrast to the configuration files for sinks and receivers, the collectors c
* [`topprocs`](./topprocsMetric.md)
* [`nfs3stat`](./nfs3Metric.md)
* [`nfs4stat`](./nfs4Metric.md)
* [`nfsiostat`](./nfsiostatMetric.md)
* [`cpufreq`](./cpufreqMetric.md)
* [`cpufreq_cpuinfo`](./cpufreqCpuinfoMetric.md)
* [`schedstat`](./schedstatMetric.md)
* [`numastats`](./numastatsMetric.md)
* [`gpfs`](./gpfsMetric.md)
* [`beegfs_meta`](./beegfsmetaMetric.md)
@@ -63,7 +51,7 @@ A collector reads data from any source, parses it to metrics and submits these m
* `Name() string`: Return the name of the collector
* `Init(config json.RawMessage) error`: Initializes the collector using the given collector-specific config in JSON. Check if needed files/commands exists, ...
* `Initialized() bool`: Check if a collector is successfully initialized
* `Read(duration time.Duration, output chan ccMessage.CCMessage)`: Read, parse and submit data to the `output` channel as [`CCMessage`](https://github.com/ClusterCockpit/cc-lib/blob/main/ccMessage/README.md). If the collector has to measure anything for some duration, use the provided function argument `duration`.
* `Read(duration time.Duration, output chan ccMetric.CCMetric)`: Read, parse and submit data to the `output` channel as [`CCMetric`](../internal/ccMetric/README.md). If the collector has to measure anything for some duration, use the provided function argument `duration`.
* `Close()`: Closes down the collector.
It is recommanded to call `setup()` in the `Init()` function.

View File

@@ -14,8 +14,8 @@ import (
"strings"
"time"
cclog "github.com/ClusterCockpit/cc-lib/ccLogger"
lp "github.com/ClusterCockpit/cc-lib/ccMessage"
cclog "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
)
const DEFAULT_BEEGFS_CMD = "beegfs-ctl"

View File

@@ -1,17 +1,5 @@
<!--
---
title: BeeGFS metadata metric collector
description: Collect metadata clientstats for `BeeGFS on Demand`
categories: [cc-metric-collector]
tags: ['Admin']
weight: 2
hugo_path: docs/reference/cc-metric-collector/collectors/beegfsmeta.md
---
-->
## `BeeGFS on Demand` collector
This Collector is to collect `BeeGFS on Demand` (BeeOND) metadata clientstats.
This Collector is to collect BeeGFS on Demand (BeeOND) metadata clientstats.
```json
"beegfs_meta": {

View File

@@ -14,8 +14,8 @@ import (
"strings"
"time"
cclog "github.com/ClusterCockpit/cc-lib/ccLogger"
lp "github.com/ClusterCockpit/cc-lib/ccMessage"
cclog "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
)
// Struct for the collector-specific JSON config

View File

@@ -1,14 +1,3 @@
<!--
---
title: "BeeGFS on Demand metric collector"
description: Collect performance metrics for BeeGFS filesystems
categories: [cc-metric-collector]
tags: ['Admin']
weight: 2
hugo_path: docs/reference/cc-metric-collector/collectors/beegfsstorage.md
---
-->
## `BeeGFS on Demand` collector
This Collector is to collect BeeGFS on Demand (BeeOND) storage stats.

View File

@@ -2,11 +2,12 @@ package collectors
import (
"encoding/json"
"os"
"sync"
"time"
cclog "github.com/ClusterCockpit/cc-lib/ccLogger"
lp "github.com/ClusterCockpit/cc-lib/ccMessage"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
cclog "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
mct "github.com/ClusterCockpit/cc-metric-collector/pkg/multiChanTicker"
)
@@ -58,7 +59,7 @@ type collectorManager struct {
// Metric collector manager access functions
type CollectorManager interface {
Init(ticker mct.MultiChanTicker, duration time.Duration, wg *sync.WaitGroup, collectConfig json.RawMessage) error
Init(ticker mct.MultiChanTicker, duration time.Duration, wg *sync.WaitGroup, collectConfigFile string) error
AddOutput(output chan lp.CCMessage)
Start()
Close()
@@ -71,7 +72,7 @@ type CollectorManager interface {
// * ticker (from variable ticker)
// * configuration (read from config file in variable collectConfigFile)
// Initialization is done for all configured collectors
func (cm *collectorManager) Init(ticker mct.MultiChanTicker, duration time.Duration, wg *sync.WaitGroup, collectConfig json.RawMessage) error {
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
@@ -80,7 +81,15 @@ func (cm *collectorManager) Init(ticker mct.MultiChanTicker, duration time.Durat
cm.ticker = ticker
cm.duration = duration
err := json.Unmarshal(collectConfig, &cm.config)
// Read collector config file
configFile, err := os.Open(collectConfigFile)
if err != nil {
cclog.Error(err.Error())
return err
}
defer configFile.Close()
jsonParser := json.NewDecoder(configFile)
err = jsonParser.Decode(&cm.config)
if err != nil {
cclog.Error(err.Error())
return err
@@ -191,9 +200,9 @@ func (cm *collectorManager) Close() {
}
// New creates a new initialized metric collector manager
func New(ticker mct.MultiChanTicker, duration time.Duration, wg *sync.WaitGroup, collectConfig json.RawMessage) (CollectorManager, error) {
func New(ticker mct.MultiChanTicker, duration time.Duration, wg *sync.WaitGroup, collectConfigFile string) (CollectorManager, error) {
cm := new(collectorManager)
err := cm.Init(ticker, duration, wg, collectConfig)
err := cm.Init(ticker, duration, wg, collectConfigFile)
if err != nil {
return nil, err
}

View File

@@ -10,8 +10,8 @@ import (
"strings"
"time"
cclog "github.com/ClusterCockpit/cc-lib/ccLogger"
lp "github.com/ClusterCockpit/cc-lib/ccMessage"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
cclog "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
)
// CPUFreqCollector

View File

@@ -1,14 +1,3 @@
<!--
---
title: CPU frequency metric collector through cpuinfo
description: Collect the CPU frequency from `/proc/cpuinfo`
categories: [cc-metric-collector]
tags: ['Admin']
weight: 2
hugo_path: docs/reference/cc-metric-collector/collectors/cpufreq_cpuinfo.md
---
-->
## `cpufreq_cpuinfo` collector
```json

View File

@@ -9,8 +9,8 @@ import (
"strings"
"time"
cclog "github.com/ClusterCockpit/cc-lib/ccLogger"
lp "github.com/ClusterCockpit/cc-lib/ccMessage"
cclog "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
"github.com/ClusterCockpit/cc-metric-collector/pkg/ccTopology"
"golang.org/x/sys/unix"
)

View File

@@ -1,14 +1,3 @@
<!--
---
title: CPU frequency metric collector through sysfs
description: Collect the CPU frequency metrics from `/sys/.../cpu/.../cpufreq`
categories: [cc-metric-collector]
tags: ['Admin']
weight: 2
hugo_path: docs/reference/cc-metric-collector/collectors/cpufreq.md
---
-->
## `cpufreq_cpuinfo` collector
```json

View File

@@ -9,8 +9,8 @@ import (
"strings"
"time"
cclog "github.com/ClusterCockpit/cc-lib/ccLogger"
lp "github.com/ClusterCockpit/cc-lib/ccMessage"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
cclog "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
sysconf "github.com/tklauser/go-sysconf"
)

View File

@@ -1,14 +1,3 @@
<!--
---
title: CPU usage metric collector
description: Collect CPU metrics from `/proc/stat`
categories: [cc-metric-collector]
tags: ['Admin']
weight: 2
hugo_path: docs/reference/cc-metric-collector/collectors/cpustat.md
---
-->
## `cpustat` collector

View File

@@ -9,7 +9,7 @@ import (
"strings"
"time"
lp "github.com/ClusterCockpit/cc-lib/ccMessage"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
influx "github.com/influxdata/line-protocol"
)

View File

@@ -1,13 +1,3 @@
<!--
---
title: CustomCommand metric collector
description: Collect messages from custom command or files
categories: [cc-metric-collector]
tags: ['Admin']
weight: 2
hugo_path: docs/reference/cc-metric-collector/collectors/customcmd.md
---
-->
## `customcmd` collector

View File

@@ -8,8 +8,8 @@ import (
"syscall"
"time"
cclog "github.com/ClusterCockpit/cc-lib/ccLogger"
lp "github.com/ClusterCockpit/cc-lib/ccMessage"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
cclog "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
)
const MOUNTFILE = `/proc/self/mounts`

View File

@@ -1,13 +1,3 @@
<!--
---
title: Disk usage statistics metric collector
description: Collect metrics for various filesystems from `/proc/self/mounts`
categories: [cc-metric-collector]
tags: ['Admin']
weight: 2
hugo_path: docs/reference/cc-metric-collector/collectors/diskstat.md
---
-->
## `diskstat` collector

View File

@@ -13,8 +13,8 @@ import (
"strings"
"time"
cclog "github.com/ClusterCockpit/cc-lib/ccLogger"
lp "github.com/ClusterCockpit/cc-lib/ccMessage"
cclog "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
)
const DEFAULT_GPFS_CMD = "mmpmon"

View File

@@ -1,14 +1,3 @@
<!--
---
title: GPFS collector
description: Collect infos about GPFS filesystems
categories: [cc-metric-collector]
tags: ['Admin']
weight: 2
hugo_path: docs/reference/cc-metric-collector/collectors/gpfs.md
---
-->
## `gpfs` collector
```json

View File

@@ -4,8 +4,8 @@ import (
"fmt"
"os"
cclog "github.com/ClusterCockpit/cc-lib/ccLogger"
lp "github.com/ClusterCockpit/cc-lib/ccMessage"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
cclog "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
"golang.org/x/sys/unix"
"encoding/json"

View File

@@ -1,13 +1,3 @@
<!--
---
title: InfiniBand Metric collector
description: Collect metrics for InfiniBand devices
categories: [cc-metric-collector]
tags: ['Admin']
weight: 2
hugo_path: docs/reference/cc-metric-collector/collectors/infiniband.md
---
-->
## `ibstat` collector

View File

@@ -9,8 +9,8 @@ import (
"strings"
"time"
cclog "github.com/ClusterCockpit/cc-lib/ccLogger"
lp "github.com/ClusterCockpit/cc-lib/ccMessage"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
cclog "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
)
// Konstante für den Pfad zu /proc/diskstats

View File

@@ -1,13 +1,3 @@
<!--
---
title: IOStat Metric collector
description: Collect metrics from `/proc/diskstats`
categories: [cc-metric-collector]
tags: ['Admin']
weight: 2
hugo_path: docs/reference/cc-metric-collector/collectors/iostat.md
---
-->
## `iostat` collector

View File

@@ -13,8 +13,8 @@ import (
"strings"
"time"
cclog "github.com/ClusterCockpit/cc-lib/ccLogger"
lp "github.com/ClusterCockpit/cc-lib/ccMessage"
cclog "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
)
const IPMISENSORS_PATH = `ipmi-sensors`

View File

@@ -1,13 +1,3 @@
<!--
---
title: IPMI Metric collector
description: Collect metrics using ipmitool or ipmi-sensors
categories: [cc-metric-collector]
tags: ['Admin']
weight: 2
hugo_path: docs/reference/cc-metric-collector/collectors/ipmi.md
---
-->
## `ipmistat` collector

View File

@@ -24,9 +24,9 @@ import (
"time"
"unsafe"
cclog "github.com/ClusterCockpit/cc-lib/ccLogger"
lp "github.com/ClusterCockpit/cc-lib/ccMessage"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
agg "github.com/ClusterCockpit/cc-metric-collector/internal/metricAggregator"
cclog "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
topo "github.com/ClusterCockpit/cc-metric-collector/pkg/ccTopology"
"github.com/NVIDIA/go-nvml/pkg/dl"
"github.com/fsnotify/fsnotify"
@@ -190,8 +190,12 @@ func getBaseFreq() float64 {
}
if math.IsNaN(freq) {
C.timer_init()
freq = float64(C.timer_getCycleClock()) / 1e3
C.power_init(0)
info := C.get_powerInfo()
if float64(info.baseFrequency) != 0 {
freq = float64(info.baseFrequency)
}
C.power_finalize()
}
return freq * 1e3
}

View File

@@ -1,13 +1,3 @@
<!--
---
title: LIKWID collector
description: Collect hardware performance events and metrics using LIKWID
categories: [cc-metric-collector]
tags: ['Admin']
weight: 2
hugo_path: docs/reference/cc-metric-collector/collectors/likwid.md
---
-->
## `likwid` collector

View File

@@ -8,8 +8,8 @@ import (
"strings"
"time"
cclog "github.com/ClusterCockpit/cc-lib/ccLogger"
lp "github.com/ClusterCockpit/cc-lib/ccMessage"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
cclog "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
)
// LoadavgCollector collects:

View File

@@ -1,14 +1,3 @@
<!--
---
title: Load average metric collector
description: Collect metrics from `/proc/loadavg`
categories: [cc-metric-collector]
tags: ['Admin']
weight: 2
hugo_path: docs/reference/cc-metric-collector/collectors/loadavg.md
---
-->
## `loadavg` collector

View File

@@ -10,8 +10,8 @@ import (
"strings"
"time"
cclog "github.com/ClusterCockpit/cc-lib/ccLogger"
lp "github.com/ClusterCockpit/cc-lib/ccMessage"
cclog "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
)
const LUSTRE_SYSFS = `/sys/fs/lustre`

View File

@@ -1,14 +1,3 @@
<!--
---
title: Lustre filesystem metric collector
description: Collect metrics for Lustre filesystems
categories: [cc-metric-collector]
tags: ['Admin']
weight: 2
hugo_path: docs/reference/cc-metric-collector/collectors/lustre.md
---
-->
## `lustrestat` collector

View File

@@ -12,8 +12,8 @@ import (
"strings"
"time"
cclog "github.com/ClusterCockpit/cc-lib/ccLogger"
lp "github.com/ClusterCockpit/cc-lib/ccMessage"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
cclog "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
)
const MEMSTATFILE = "/proc/meminfo"

View File

@@ -1,14 +1,3 @@
<!--
---
title: Memory statistics metric collector
description: Collect metrics from `/proc/meminfo`
categories: [cc-metric-collector]
tags: ['Admin']
weight: 2
hugo_path: docs/reference/cc-metric-collector/collectors/memstat.md
---
-->
## `memstat` collector

View File

@@ -5,7 +5,7 @@ import (
"fmt"
"time"
lp "github.com/ClusterCockpit/cc-lib/ccMessage"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
)
type MetricCollector interface {
@@ -14,7 +14,7 @@ type MetricCollector interface {
Initialized() bool // Is metric collector initialized?
Parallel() bool
Read(duration time.Duration, output chan lp.CCMessage) // Read metrics from metric collector
Close() // Close / finish metric collector
Close() // Close / finish metric collector
}
type metricCollector struct {

View File

@@ -9,8 +9,8 @@ import (
"strings"
"time"
cclog "github.com/ClusterCockpit/cc-lib/ccLogger"
lp "github.com/ClusterCockpit/cc-lib/ccMessage"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
cclog "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
)
const NETSTATFILE = "/proc/net/dev"

View File

@@ -1,13 +1,3 @@
<!--
---
title: Network device metric collector
description: Collect metrics for network devices through procfs
categories: [cc-metric-collector]
tags: ['Admin']
weight: 2
hugo_path: docs/reference/cc-metric-collector/collectors/netstat.md
---
-->
## `netstat` collector

View File

@@ -1,14 +1,3 @@
<!--
---
title: NFS network filesystem (v3) metric collector
description: Collect metrics for NFS network filesystems in version 3
categories: [cc-metric-collector]
tags: ['Admin']
weight: 2
hugo_path: docs/reference/cc-metric-collector/collectors/nfs3.md
---
-->
## `nfs3stat` collector

View File

@@ -1,14 +1,3 @@
<!--
---
title: NFS network filesystem (v4) metric collector
description: Collect metrics for NFS network filesystems in version 4
categories: [cc-metric-collector]
tags: ['Admin']
weight: 2
hugo_path: docs/reference/cc-metric-collector/collectors/nfs4.md
---
-->
## `nfs4stat` collector

View File

@@ -11,7 +11,7 @@ import (
"strings"
"time"
lp "github.com/ClusterCockpit/cc-lib/ccMessage"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
)
// First part contains the code for the general NfsCollector.

View File

@@ -9,8 +9,8 @@ import (
"strings"
"time"
cclog "github.com/ClusterCockpit/cc-lib/ccLogger"
lp "github.com/ClusterCockpit/cc-lib/ccMessage"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
cclog "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
)
// These are the fields we read from the JSON configuration
@@ -171,7 +171,7 @@ func (m *NfsIOStatCollector) Read(interval time.Duration, output chan lp.CCMessa
}
}
if !found {
delete(m.data, mntpoint)
m.data[mntpoint] = nil
}
}
}

View File

@@ -1,14 +1,3 @@
<!--
---
title: NFS network filesystem metrics from procfs
description: Collect NFS network filesystem metrics for mounts from `/proc/self/mountstats`
categories: [cc-metric-collector]
tags: ['Admin']
weight: 2
hugo_path: docs/reference/cc-metric-collector/collectors/nfsio.md
---
-->
## `nfsiostat` collector
```json

View File

@@ -10,8 +10,8 @@ import (
"strings"
"time"
cclog "github.com/ClusterCockpit/cc-lib/ccLogger"
lp "github.com/ClusterCockpit/cc-lib/ccMessage"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
cclog "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
)
type NUMAStatsCollectorConfig struct {

View File

@@ -1,13 +1,3 @@
<!--
---
title: NUMAStat collector
description: Collect infos about NUMA domains
categories: [cc-metric-collector]
tags: ['Admin']
weight: 2
hugo_path: docs/reference/cc-metric-collector/collectors/numastat.md
---
-->
## `numastat` collector

View File

@@ -8,8 +8,8 @@ import (
"strings"
"time"
cclog "github.com/ClusterCockpit/cc-lib/ccLogger"
lp "github.com/ClusterCockpit/cc-lib/ccMessage"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
cclog "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
"github.com/NVIDIA/go-nvml/pkg/nvml"
)

View File

@@ -1,13 +1,3 @@
<!--
---
title: "Nvidia NVML metric collector"
description: Collect metrics for Nvidia GPUs using the NVML
categories: [cc-metric-collector]
tags: ['Admin']
weight: 2
hugo_path: docs/reference/cc-metric-collector/collectors/nvidia.md
---
-->
## `nvidia` collector

View File

@@ -9,20 +9,29 @@ import (
"strings"
"time"
cclog "github.com/ClusterCockpit/cc-lib/ccLogger"
lp "github.com/ClusterCockpit/cc-lib/ccMessage"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
cclog "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
)
// running average power limit (RAPL) monitoring attributes for a zone
// Only for Intel systems
type RAPLZoneInfo struct {
energy int64 // current reading of the energy counter in micro joules
maxEnergyRange int64 // Range of the above energy counter in micro-joules
energyTimestamp time.Time // timestamp when energy counter was read
energyFilepath string // path to a file containing the zones current energy counter in micro joules
shortTermFilepath string // path to short term power limit
longTermFilepath string // path to long term power limit
enabledFilepath string // path to check whether limits are enabled
name string
// tags describing the RAPL zone:
// * zone_name, subzone_name: e.g. psys, dram, core, uncore, package-0
// * zone_id: e.g. 0:1 (zone 0 sub zone 1)
tags map[string]string
energyFilepath string // path to a file containing the zones current energy counter in micro joules
energy int64 // current reading of the energy counter in micro joules
energyTimestamp time.Time // timestamp when energy counter was read
maxEnergyRange int64 // Range of the above energy counter in micro-joules
// type=socket for dram, core, uncore, package-* and type=node for psys
// type-id=socket id
tags map[string]string
}
type RAPLCollector struct {
@@ -33,12 +42,40 @@ type RAPLCollector struct {
// * 0:1 for zone 0 subzone 1
ExcludeByID []string `json:"exclude_device_by_id,omitempty"`
// Exclude names for RAPL zones, e.g. psys, dram, core, uncore, package-0
ExcludeByName []string `json:"exclude_device_by_name,omitempty"`
ExcludeByName []string `json:"exclude_device_by_name,omitempty"`
SkipEnergyReading bool `json:"skip_energy_reading,omitempty"`
SkipLimitsReading bool `json:"skip_limits_reading,omitempty"`
OnlyEnabledLimits bool `json:"only_enabled_limits,omitempty"`
}
RAPLZoneInfo []RAPLZoneInfo
raplZoneInfo []RAPLZoneInfo
meta map[string]string // default meta information
}
// Get the path to the power limit file for zone selectable by limit name
// Common limit names for Intel systems are
// - long_term
// - short_term
// Does not support AMD as AMD systems do not provide the power limits
// through sysfs
func ZoneLimitFile(folder string, limit_name string) string {
nameGlob := filepath.Join(folder, "constraint_*_name")
candidates, err := filepath.Glob(nameGlob)
if err == nil {
for _, c := range candidates {
if v, err := os.ReadFile(c); err == nil {
if strings.TrimSpace(string(v)) == limit_name {
var i int
n, err := fmt.Sscanf(filepath.Base(c), "constraint_%d_name", &i)
if err == nil && n == 1 {
return filepath.Join(folder, fmt.Sprintf("constraint_%d_power_limit_uw", i))
}
}
}
}
}
return ""
}
// Init initializes the running average power limit (RAPL) collector
func (m *RAPLCollector) Init(config json.RawMessage) error {
@@ -58,6 +95,9 @@ func (m *RAPLCollector) Init(config json.RawMessage) error {
}
// Read in the JSON configuration
m.config.SkipEnergyReading = false
m.config.SkipLimitsReading = false
m.config.OnlyEnabledLimits = true
if len(config) > 0 {
err = json.Unmarshal(config, &m.config)
if err != nil {
@@ -83,50 +123,62 @@ func (m *RAPLCollector) Init(config json.RawMessage) error {
// readZoneInfo reads RAPL monitoring attributes for a zone given by zonePath
// See: https://www.kernel.org/doc/html/latest/power/powercap/powercap.html#monitoring-attributes
readZoneInfo := func(zonePath string) (z struct {
name string // zones name e.g. psys, dram, core, uncore, package-0
energyFilepath string // path to a file containing the zones current energy counter in micro joules
energy int64 // current reading of the energy counter in micro joules
energyTimestamp time.Time // timestamp when energy counter was read
maxEnergyRange int64 // Range of the above energy counter in micro-joules
ok bool // Are all information available?
name string // zones name e.g. psys, dram, core, uncore, package-0
energyFilepath string // path to a file containing the zones current energy counter in micro joules
energy int64 // current reading of the energy counter in micro joules
energyTimestamp time.Time // timestamp when energy counter was read
maxEnergyRange int64 // Range of the above energy counter in micro-joules
shortTermFilepath string
longTermFilepath string
enabledFilepath string
}) {
// zones name e.g. psys, dram, core, uncore, package-0
foundName := false
if v, err :=
os.ReadFile(
filepath.Join(zonePath, "name")); err == nil {
foundName = true
z.name = strings.TrimSpace(string(v))
}
// path to a file containing the zones current energy counter in micro joules
z.energyFilepath = filepath.Join(zonePath, "energy_uj")
// current reading of the energy counter in micro joules
foundEnergy := false
if v, err := os.ReadFile(z.energyFilepath); err == nil {
// timestamp when energy counter was read
z.energyTimestamp = time.Now()
if i, err := strconv.ParseInt(strings.TrimSpace(string(v)), 10, 64); err == nil {
foundEnergy = true
z.energy = i
if os.Geteuid() == 0 && (!m.config.SkipEnergyReading) {
// path to a file containing the zones current energy counter in micro joules
z.energyFilepath = filepath.Join(zonePath, "energy_uj")
// current reading of the energy counter in micro joules
if v, err := os.ReadFile(z.energyFilepath); err == nil {
if i, err := strconv.ParseInt(strings.TrimSpace(string(v)), 10, 64); err == nil {
z.energy = i
// timestamp when energy counter was read
z.energyTimestamp = time.Now()
}
} else {
cclog.ComponentError(m.name, "Cannot read energy file for ", z.name, ":", err.Error())
}
// Range of the above energy counter in micro-joules
if v, err :=
os.ReadFile(
filepath.Join(zonePath, "max_energy_range_uj")); err == nil {
if i, err := strconv.ParseInt(strings.TrimSpace(string(v)), 10, 64); err == nil {
z.maxEnergyRange = i
}
}
} else {
cclog.ComponentInfo(m.name, "Energy readings for", zonePath, "disabled")
}
// Range of the above energy counter in micro-joules
foundMaxEnergyRange := false
if v, err :=
os.ReadFile(
filepath.Join(zonePath, "max_energy_range_uj")); err == nil {
if i, err := strconv.ParseInt(strings.TrimSpace(string(v)), 10, 64); err == nil {
foundMaxEnergyRange = true
z.maxEnergyRange = i
if !m.config.SkipLimitsReading {
z.shortTermFilepath = ZoneLimitFile(zonePath, "short_term")
if _, err := os.Stat(z.shortTermFilepath); err != nil {
z.shortTermFilepath = ""
}
z.longTermFilepath = ZoneLimitFile(zonePath, "long_term")
if _, err := os.Stat(z.longTermFilepath); err != nil {
z.longTermFilepath = ""
}
z.enabledFilepath = filepath.Join(zonePath, "enabled")
} else {
cclog.ComponentInfo(m.name, "Power limit readings for", zonePath, "disabled")
}
// Are all information available?
z.ok = foundName && foundEnergy && foundMaxEnergyRange
return
}
@@ -143,25 +195,42 @@ func (m *RAPLCollector) Init(config json.RawMessage) error {
for _, zonePath := range zonesPath {
zoneID := strings.TrimPrefix(zonePath, zonePrefix)
zonetags := make(map[string]string)
z := readZoneInfo(zonePath)
if z.ok &&
!isIDExcluded[zoneID] &&
if !isIDExcluded[zoneID] &&
!isNameExcluded[z.name] {
si := RAPLZoneInfo{
tags: make(map[string]string),
energyFilepath: z.energyFilepath,
energy: z.energy,
energyTimestamp: z.energyTimestamp,
maxEnergyRange: z.maxEnergyRange,
shortTermFilepath: z.shortTermFilepath,
longTermFilepath: z.longTermFilepath,
enabledFilepath: z.enabledFilepath,
name: z.name,
}
si.tags["type"] = "node"
si.tags["type-id"] = "0"
var pid int = 0
if strings.HasPrefix(z.name, "package-") {
n, err := fmt.Sscanf(z.name, "package-%d", &pid)
if err == nil && n == 1 {
si.tags["type-id"] = fmt.Sprintf("%d", pid)
si.tags["type"] = "socket"
}
si.name = "pkg"
}
// Add RAPL monitoring attributes for a zone
m.RAPLZoneInfo =
append(
m.RAPLZoneInfo,
RAPLZoneInfo{
tags: map[string]string{
"id": zoneID,
"zone_name": z.name,
},
energyFilepath: z.energyFilepath,
energy: z.energy,
energyTimestamp: z.energyTimestamp,
maxEnergyRange: z.maxEnergyRange,
})
if _, ok1 := si.tags["type"]; ok1 {
if _, ok2 := si.tags["type-id"]; ok2 {
m.raplZoneInfo = append(m.raplZoneInfo, si)
zonetags["type"] = si.tags["type"]
zonetags["type-id"] = si.tags["type-id"]
}
}
}
// find all sub zones for the given zone
@@ -174,29 +243,32 @@ func (m *RAPLCollector) Init(config json.RawMessage) error {
for _, subZonePath := range subZonesPath {
subZoneID := strings.TrimPrefix(subZonePath, subZonePrefix)
sz := readZoneInfo(subZonePath)
if len(zoneID) > 0 && len(z.name) > 0 &&
sz.ok &&
!isIDExcluded[zoneID+":"+subZoneID] &&
!isNameExcluded[sz.name] {
m.RAPLZoneInfo =
append(
m.RAPLZoneInfo,
RAPLZoneInfo{
tags: map[string]string{
"id": zoneID + ":" + subZoneID,
"zone_name": z.name,
"sub_zone_name": sz.name,
},
energyFilepath: sz.energyFilepath,
energy: sz.energy,
energyTimestamp: sz.energyTimestamp,
maxEnergyRange: sz.maxEnergyRange,
})
si := RAPLZoneInfo{
tags: zonetags,
energyFilepath: sz.energyFilepath,
energy: sz.energy,
energyTimestamp: sz.energyTimestamp,
maxEnergyRange: sz.maxEnergyRange,
shortTermFilepath: sz.shortTermFilepath,
longTermFilepath: sz.longTermFilepath,
enabledFilepath: sz.enabledFilepath,
name: sz.name,
}
if _, ok1 := si.tags["type"]; ok1 {
if _, ok2 := si.tags["type-id"]; ok2 {
m.raplZoneInfo = append(m.raplZoneInfo, si)
}
}
}
}
}
if m.RAPLZoneInfo == nil {
if m.raplZoneInfo == nil {
return fmt.Errorf("no running average power limit (RAPL) device found in %s", controlTypePath)
}
@@ -205,7 +277,7 @@ func (m *RAPLCollector) Init(config json.RawMessage) error {
cclog.ComponentDebug(
m.name,
"initialized",
len(m.RAPLZoneInfo),
len(m.raplZoneInfo),
"zones with running average power limit (RAPL) monitoring attributes")
m.init = true
@@ -216,40 +288,89 @@ func (m *RAPLCollector) Init(config json.RawMessage) error {
// See: https://www.kernel.org/doc/html/latest/power/powercap/powercap.html#monitoring-attributes
func (m *RAPLCollector) Read(interval time.Duration, output chan lp.CCMessage) {
for i := range m.RAPLZoneInfo {
p := &m.RAPLZoneInfo[i]
for i := range m.raplZoneInfo {
p := &m.raplZoneInfo[i]
// Read current value of the energy counter in micro joules
if v, err := os.ReadFile(p.energyFilepath); err == nil {
energyTimestamp := time.Now()
if i, err := strconv.ParseInt(strings.TrimSpace(string(v)), 10, 64); err == nil {
energy := i
if os.Geteuid() == 0 && (!m.config.SkipEnergyReading) {
// Read current value of the energy counter in micro joules
if v, err := os.ReadFile(p.energyFilepath); err == nil {
energyTimestamp := time.Now()
if i, err := strconv.ParseInt(strings.TrimSpace(string(v)), 10, 64); err == nil {
energy := i
// Compute average power (Δ energy / Δ time)
energyDiff := energy - p.energy
if energyDiff < 0 {
// Handle overflow:
// ( p.maxEnergyRange - p.energy ) + energy
// = p.maxEnergyRange + ( energy - p.energy )
// = p.maxEnergyRange + diffEnergy
energyDiff += p.maxEnergyRange
// Compute average power (Δ energy / Δ time)
energyDiff := energy - p.energy
if energyDiff < 0 {
// Handle overflow:
// ( p.maxEnergyRange - p.energy ) + energy
// = p.maxEnergyRange + ( energy - p.energy )
// = p.maxEnergyRange + diffEnergy
energyDiff += p.maxEnergyRange
}
timeDiff := energyTimestamp.Sub(p.energyTimestamp)
averagePower := float64(energyDiff) / float64(timeDiff.Microseconds())
y, err := lp.NewMetric(
fmt.Sprintf("rapl_%s_average_power", p.name),
p.tags,
m.meta,
averagePower,
energyTimestamp)
if err == nil {
output <- y
}
e, err := lp.NewMetric(
fmt.Sprintf("rapl_%s_energy", p.name),
p.tags,
m.meta,
float64(energyDiff)*1e-3,
energyTimestamp)
if err == nil {
e.AddMeta("unit", "Joules")
output <- e
}
// Save current energy counter state
p.energy = energy
p.energyTimestamp = energyTimestamp
}
timeDiff := energyTimestamp.Sub(p.energyTimestamp)
averagePower := float64(energyDiff) / float64(timeDiff.Microseconds())
y, err := lp.NewMessage(
"rapl_average_power",
p.tags,
m.meta,
map[string]interface{}{"value": averagePower},
energyTimestamp)
if err == nil {
output <- y
}
}
// https://www.kernel.org/doc/html/latest/power/powercap/powercap.html#constraints
if !m.config.SkipLimitsReading {
skip := false
if m.config.OnlyEnabledLimits {
if v, err := os.ReadFile(p.enabledFilepath); err == nil {
if strings.TrimSpace(string(v)) == "0" {
skip = true
}
}
}
if !skip {
if len(p.shortTermFilepath) > 0 {
if v, err := os.ReadFile(p.shortTermFilepath); err == nil {
if i, err := strconv.ParseInt(strings.TrimSpace(string(v)), 10, 64); err == nil {
name := fmt.Sprintf("rapl_%s_limit_short_term", p.name)
y, err := lp.NewMetric(name, p.tags, m.meta, i/1e6, time.Now())
if err == nil {
output <- y
}
}
}
}
// Save current energy counter state
p.energy = energy
p.energyTimestamp = energyTimestamp
if len(p.longTermFilepath) > 0 {
if v, err := os.ReadFile(p.longTermFilepath); err == nil {
if i, err := strconv.ParseInt(strings.TrimSpace(string(v)), 10, 64); err == nil {
name := fmt.Sprintf("rapl_%s_limit_long_term", p.name)
y, err := lp.NewMetric(name, p.tags, m.meta, i/1e6, time.Now())
if err == nil {
output <- y
}
}
}
}
}
}
}

View File

@@ -1,26 +1,23 @@
<!--
---
title: RAPL metric collector
description: Collect energy data through the RAPL sysfs interface
categories: [cc-metric-collector]
tags: ['Admin']
weight: 2
hugo_path: docs/reference/cc-metric-collector/collectors/rapl.md
---
-->
## `rapl` collector
This collector reads running average power limit (RAPL) monitoring attributes to compute average power consumption metrics. See <https://www.kernel.org/doc/html/latest/power/powercap/powercap.html#monitoring-attributes>.
The Likwid metric collector provides similar functionality.
This collector reads running average power limit (RAPL) monitoring attributes to compute average power consumption metrics. See <https://www.kernel.org/doc/html/latest/power/powercap/powercap.html>.
```json
"rapl": {
"exclude_device_by_id": ["0:1", "0:2"],
"exclude_device_by_name": ["psys"]
"exclude_device_by_name": ["psys"],
"skip_energy_reading": false,
"skip_limits_reading": false,
"only_enabled_limits": true
}
```
Metrics:
* `rapl_average_power`: average power consumption in Watt. The average is computed over the entire runtime from the last measurement to the current measurement
* `rapl_<domain>_average_power`: average power consumption in Watt. The average is computed over the entire runtime from the last measurement to the current measurement
* `rapl_<domain>_energy`: Difference from the last measurement
* `rapl_<domain>_limit_short_term`: Short term powercap setting for the domain
* `rapl_<domain>_limit_long_term`: Long term powercap setting for the domain
Only the `rapl_<domain>_average_power` and `rapl_<domain>_energy` metrics require root-permissions. The limits can be read as user. Some domains have limits available but they are not enabled. By default, only enabled domain limits are collected.
Energy and power measurments can also be done with the Likwid metric collector.

View File

@@ -6,8 +6,8 @@ import (
"fmt"
"time"
cclog "github.com/ClusterCockpit/cc-lib/ccLogger"
lp "github.com/ClusterCockpit/cc-lib/ccMessage"
cclog "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
"github.com/ClusterCockpit/go-rocm-smi/pkg/rocm_smi"
)

View File

@@ -1,14 +1,3 @@
<!--
---
title: "ROCm SMI metric collector"
description: Collect metrics for AMD GPUs using the SMI library
categories: [cc-metric-collector]
tags: ['Admin']
weight: 2
hugo_path: docs/reference/cc-metric-collector/collectors/rocmsmi.md
---
-->
## `rocm_smi` collector

View File

@@ -4,8 +4,8 @@ import (
"encoding/json"
"time"
cclog "github.com/ClusterCockpit/cc-lib/ccLogger"
lp "github.com/ClusterCockpit/cc-lib/ccMessage"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
cclog "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
)
// These are the fields we read from the JSON configuration

View File

@@ -5,8 +5,8 @@ import (
"sync"
"time"
cclog "github.com/ClusterCockpit/cc-lib/ccLogger"
lp "github.com/ClusterCockpit/cc-lib/ccMessage"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
cclog "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
)
// These are the fields we read from the JSON configuration

View File

@@ -10,8 +10,8 @@ import (
"strings"
"time"
cclog "github.com/ClusterCockpit/cc-lib/ccLogger"
lp "github.com/ClusterCockpit/cc-lib/ccMessage"
cclog "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
)
const SCHEDSTATFILE = `/proc/schedstat`

View File

@@ -1,13 +1,3 @@
<!--
---
title: SchedStat Metric collector
description: Collect metrics from `/proc/schedstat`
categories: [cc-metric-collector]
tags: ['Admin']
weight: 2
hugo_path: docs/reference/cc-metric-collector/collectors/schedstat.md
---
-->
## `schedstat` collector
```json

View File

@@ -6,8 +6,8 @@ import (
"syscall"
"time"
cclog "github.com/ClusterCockpit/cc-lib/ccLogger"
lp "github.com/ClusterCockpit/cc-lib/ccMessage"
cclog "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
)
type SelfCollectorConfig struct {

View File

@@ -1,14 +1,3 @@
<!--
---
title: Self-monitoring metric collector
description: Collect metrics from the execution of cc-metric-collector itself
categories: [cc-metric-collector]
tags: ['Admin']
weight: 2
hugo_path: docs/reference/cc-metric-collector/collectors/self.md
---
-->
## `self` collector
```json

View File

@@ -9,8 +9,8 @@ import (
"strings"
"time"
cclog "github.com/ClusterCockpit/cc-lib/ccLogger"
lp "github.com/ClusterCockpit/cc-lib/ccMessage"
cclog "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
)
// See: https://www.kernel.org/doc/html/latest/hwmon/sysfs-interface.html

View File

@@ -1,14 +1,3 @@
<!--
---
title: Temperature metric collector
description: Collect thermal metrics from `/sys/class/hwmon/*`
categories: [cc-metric-collector]
tags: ['Admin']
weight: 2
hugo_path: docs/reference/cc-metric-collector/collectors/temp.md
---
-->
## `tempstat` collector

View File

@@ -9,7 +9,7 @@ import (
"strings"
"time"
lp "github.com/ClusterCockpit/cc-lib/ccMessage"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
)
const MAX_NUM_PROCS = 10

View File

@@ -1,15 +1,3 @@
<!--
---
title: TopProcs collector
description: Collect infos about most CPU-consuming processes
categories: [cc-metric-collector]
tags: ['Admin']
weight: 2
hugo_path: docs/reference/cc-metric-collector/collectors/topprocs.md
---
-->
## `topprocs` collector

View File

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

17
go.mod
View File

@@ -1,14 +1,17 @@
module github.com/ClusterCockpit/cc-metric-collector
go 1.23.4
go 1.23.0
toolchain go1.23.7
toolchain go1.23.2
require (
github.com/ClusterCockpit/cc-lib v0.1.1
github.com/ClusterCockpit/cc-energy-manager v0.0.0-20240709142550-dd446f7ab900
github.com/ClusterCockpit/cc-lib v0.1.0-beta.1
github.com/ClusterCockpit/cc-units v0.4.0
github.com/ClusterCockpit/go-rocm-smi v0.3.0
github.com/NVIDIA/go-nvml v0.12.0-2
github.com/PaesslerAG/gval v1.2.2
github.com/expr-lang/expr v1.16.9
github.com/fsnotify/fsnotify v1.7.0
github.com/gorilla/mux v1.8.1
github.com/influxdata/influxdb-client-go/v2 v2.14.0
@@ -20,16 +23,14 @@ require (
github.com/tklauser/go-sysconf v0.3.13
golang.design/x/thread v0.0.0-20210122121316-335e9adffdf1
golang.org/x/exp v0.0.0-20250215185904-eff6e970281f
golang.org/x/sys v0.30.0
golang.org/x/sys v0.28.0
)
require (
github.com/ClusterCockpit/cc-backend v1.4.2 // indirect
github.com/ClusterCockpit/cc-units v0.4.0 // indirect
github.com/apapsch/go-jsonmerge/v2 v2.0.0 // indirect
github.com/beorn7/perks v1.0.1 // indirect
github.com/cespare/xxhash/v2 v2.3.0 // indirect
github.com/expr-lang/expr v1.17.0 // indirect
github.com/google/uuid v1.6.0 // indirect
github.com/klauspost/compress v1.17.9 // indirect
github.com/munnerz/goautoneg v0.0.0-20191010083416-a7dc8b61c822 // indirect
@@ -42,7 +43,7 @@ require (
github.com/santhosh-tekuri/jsonschema/v5 v5.3.1 // indirect
github.com/shopspring/decimal v1.3.1 // indirect
github.com/tklauser/numcpus v0.7.0 // indirect
golang.org/x/crypto v0.35.0 // indirect
golang.org/x/net v0.36.0 // indirect
golang.org/x/crypto v0.31.0 // indirect
golang.org/x/net v0.31.0 // indirect
google.golang.org/protobuf v1.35.2 // indirect
)

12
go.sum
View File

@@ -1,9 +1,9 @@
github.com/ClusterCockpit/cc-backend v1.4.2 h1:kTOzqkh9N0564N9nqQThnSs7TAfg8RLgvSm00e5HtIc=
github.com/ClusterCockpit/cc-backend v1.4.2/go.mod h1:g8TNHXe4AXej26snu2//jO3mUF980elT93iV/k11O/c=
github.com/ClusterCockpit/cc-energy-manager v0.0.0-20240709142550-dd446f7ab900 h1:6+WNav16uWTEDC09hkZKEHfBhtc91p/ZcjgCtyntuIg=
github.com/ClusterCockpit/cc-energy-manager v0.0.0-20240709142550-dd446f7ab900/go.mod h1:EbYeC5t+Y0kW1Q1pP2n9zMqbeYEJITG8YGvAUihXVn4=
github.com/ClusterCockpit/cc-lib v0.1.0-beta.1 h1:dz9j0g2cod8+SMDjuoIY6ISpiHHeekhX6yQaeiwiwJw=
github.com/ClusterCockpit/cc-lib v0.1.0-beta.1/go.mod h1:kXMskla1i5ZSfXW0vVRIHgGeXMU5zu2PzYOYnUaOr80=
github.com/ClusterCockpit/cc-lib v0.1.1 h1:AXZWYUzgTaE/WdxLNSWPR7FJoA5WlzvYZxw4gIw3gNw=
github.com/ClusterCockpit/cc-lib v0.1.1/go.mod h1:SHKcWW/+kN+pcofAtHJFxvmx1FV0VIJuQv5PuT0HDcc=
github.com/ClusterCockpit/cc-units v0.4.0 h1:zP5DOu99GmErW0tCDf0gcLrlWt42RQ9dpoONEOh4cI0=
github.com/ClusterCockpit/cc-units v0.4.0/go.mod h1:3S3PAhAayS3pbgcT4q9Vn9VJw22Op51X0YimtG77zBw=
github.com/ClusterCockpit/go-rocm-smi v0.3.0 h1:1qZnSpG7/NyLtc7AjqnUL9Jb8xtqG1nMVgp69rJfaR8=
@@ -29,8 +29,6 @@ github.com/davecgh/go-spew v1.1.1 h1:vj9j/u1bqnvCEfJOwUhtlOARqs3+rkHYY13jYWTU97c
github.com/davecgh/go-spew v1.1.1/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
github.com/expr-lang/expr v1.16.9 h1:WUAzmR0JNI9JCiF0/ewwHB1gmcGw5wW7nWt8gc6PpCI=
github.com/expr-lang/expr v1.16.9/go.mod h1:8/vRC7+7HBzESEqt5kKpYXxrxkr31SaO8r40VO/1IT4=
github.com/expr-lang/expr v1.17.0 h1:+vpszOyzKLQXC9VF+wA8cVA0tlA984/Wabc/1hF9Whg=
github.com/expr-lang/expr v1.17.0/go.mod h1:8/vRC7+7HBzESEqt5kKpYXxrxkr31SaO8r40VO/1IT4=
github.com/frankban/quicktest v1.11.0/go.mod h1:K+q6oSqb0W0Ininfk863uOk1lMy69l/P6txr3mVT54s=
github.com/frankban/quicktest v1.11.2/go.mod h1:K+q6oSqb0W0Ininfk863uOk1lMy69l/P6txr3mVT54s=
github.com/frankban/quicktest v1.13.0 h1:yNZif1OkDfNoDfb9zZa9aXIpejNR4F23Wely0c+Qdqk=
@@ -115,19 +113,13 @@ golang.design/x/thread v0.0.0-20210122121316-335e9adffdf1 h1:P7S/GeHBAFEZIYp0ePP
golang.design/x/thread v0.0.0-20210122121316-335e9adffdf1/go.mod h1:9CWpnTUmlQkfdpdutA1nNf4iE5lAVt3QZOu0Z6hahBE=
golang.org/x/crypto v0.31.0 h1:ihbySMvVjLAeSH1IbfcRTkD/iNscyz8rGzjF/E5hV6U=
golang.org/x/crypto v0.31.0/go.mod h1:kDsLvtWBEx7MV9tJOj9bnXsPbxwJQ6csT/x4KIN4Ssk=
golang.org/x/crypto v0.35.0 h1:b15kiHdrGCHrP6LvwaQ3c03kgNhhiMgvlhxHQhmg2Xs=
golang.org/x/crypto v0.35.0/go.mod h1:dy7dXNW32cAb/6/PRuTNsix8T+vJAqvuIy5Bli/x0YQ=
golang.org/x/exp v0.0.0-20250215185904-eff6e970281f h1:oFMYAjX0867ZD2jcNiLBrI9BdpmEkvPyi5YrBGXbamg=
golang.org/x/exp v0.0.0-20250215185904-eff6e970281f/go.mod h1:BHOTPb3L19zxehTsLoJXVaTktb06DFgmdW6Wb9s8jqk=
golang.org/x/net v0.31.0 h1:68CPQngjLL0r2AlUKiSxtQFKvzRVbnzLwMUn5SzcLHo=
golang.org/x/net v0.31.0/go.mod h1:P4fl1q7dY2hnZFxEk4pPSkDHF+QqjitcnDjUQyMM+pM=
golang.org/x/net v0.36.0 h1:vWF2fRbw4qslQsQzgFqZff+BItCvGFQqKzKIzx1rmoA=
golang.org/x/net v0.36.0/go.mod h1:bFmbeoIPfrw4sMHNhb4J9f6+tPziuGjq7Jk/38fxi1I=
golang.org/x/sys v0.0.0-20210122093101-04d7465088b8/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
golang.org/x/sys v0.28.0 h1:Fksou7UEQUWlKvIdsqzJmUmCX3cZuD2+P3XyyzwMhlA=
golang.org/x/sys v0.28.0/go.mod h1:/VUhepiaJMQUp4+oa/7Zr1D23ma6VTLIYjOOTFZPUcA=
golang.org/x/sys v0.30.0 h1:QjkSwP/36a20jFYWkSue1YwXzLmsV5Gfq7Eiy72C1uc=
golang.org/x/sys v0.30.0/go.mod h1:/VUhepiaJMQUp4+oa/7Zr1D23ma6VTLIYjOOTFZPUcA=
golang.org/x/xerrors v0.0.0-20191204190536-9bdfabe68543/go.mod h1:I/5z698sn9Ka8TeJc9MKroUUfqBBauWjQqLJ2OPfmY0=
google.golang.org/protobuf v1.35.2 h1:8Ar7bF+apOIoThw1EdZl0p1oWvMqTHmpA2fRTyZO8io=
google.golang.org/protobuf v1.35.2/go.mod h1:9fA7Ob0pmnwhb644+1+CVWFRbNajQ6iRojtC/QF5bRE=

View File

@@ -1,14 +1,3 @@
<!--
---
title: Metric Aggregator
description: Subsystem for evaluating expressions on metrics (deprecated)
categories: [cc-metric-collector]
tags: ['Developer']
weight: 1
hugo_path: docs/reference/cc-metric-collector/internal/metricaggregator/_index.md
---
-->
# The MetricAggregator
In some cases, further combination of metrics or raw values is required. For that strings like `foo + 1` with runtime dependent `foo` need to be evaluated. The MetricAggregator relies on the [`gval`](https://github.com/PaesslerAG/gval) Golang package to perform all expression evaluation. The `gval` package provides the basic arithmetic operations but the MetricAggregator defines additional ones.

View File

@@ -9,9 +9,9 @@ import (
"sync"
"time"
cclog "github.com/ClusterCockpit/cc-lib/ccLogger"
cclog "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
lp "github.com/ClusterCockpit/cc-lib/ccMessage"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
topo "github.com/ClusterCockpit/cc-metric-collector/pkg/ccTopology"
"github.com/PaesslerAG/gval"

View File

@@ -1,22 +1,11 @@
<!--
---
title: Message Router
description: Routing component inside cc-metric-collector
categories: [cc-metric-collector]
tags: ['Developer']
weight: 1
hugo_path: docs/reference/cc-metric-collector/internal/metricrouter/_index.md
---
-->
# CC Metric Router
The CCMetric router sits in between the collectors and the sinks and can be used to add and remove tags to/from traversing [CCMessages](https://pkg.go.dev/github.com/ClusterCockpit/cc-lib/ccMessage).
The CCMetric router sits in between the collectors and the sinks and can be used to add and remove tags to/from traversing [CCMessages](https://pkg.go.dev/github.com/ClusterCockpit/cc-energy-manager@v0.0.0-20240919152819-92a17f2da4f7/pkg/cc-message.
# Configuration
**Note**: Use the [message processor configuration](https://github.com/ClusterCockpit/cc-lib/blob/main/messageProcessor/README.md) with option `process_messages`.
**Note**: Use the [message processor configuration](../../pkg/messageProcessor/README.md) with option `process_messages`.
```json
{
@@ -80,7 +69,7 @@ The CCMetric router sits in between the collectors and the sinks and can be used
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.
**Note**: Use the [message processor configuration](https://github.com/ClusterCockpit/cc-lib/blob/main/messageProcessor/README.md) (option `process_messages`) instead of `add_tags`, `delete_tags`, `drop_metrics`, `drop_metrics_if`, `rename_metrics`, `normalize_units` and `change_unit_prefix`. These options are deprecated and will be removed in future versions. Until then, they are added to the message processor.
**Note**: Use the [message processor configuration](../../pkg/messageProcessor/README.md) (option `process_messages`) instead of `add_tags`, `delete_tags`, `drop_metrics`, `drop_metrics_if`, `rename_metrics`, `normalize_units` and `change_unit_prefix`. These options are deprecated and will be removed in future versions. Until then, they are added to the message processor.
# Processing order in the router
@@ -274,7 +263,7 @@ The above configuration, collects all metric values for metrics evaluating `if`
If you are not interested in the input metrics `sub_metric_%d+` at all, you can add the same condition used here to the `drop_metrics_if` section to drop them.
Use cases for `interval_aggregates`:
- Combine multiple metrics of the a collector to a new one like the [MemstatCollector](../../collectors/memstatMetric.md) does it for `mem_used`:
- Combine multiple metrics of the a collector to a new one like the [MemstatCollector](../../collectors/memstatMetric.md) does it for `mem_used`)):
```json
{
"name" : "mem_used",

View File

@@ -4,10 +4,10 @@ import (
"sync"
"time"
cclog "github.com/ClusterCockpit/cc-lib/ccLogger"
cclog "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
lp "github.com/ClusterCockpit/cc-lib/ccMessage"
agg "github.com/ClusterCockpit/cc-metric-collector/internal/metricAggregator"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
mct "github.com/ClusterCockpit/cc-metric-collector/pkg/multiChanTicker"
)

View File

@@ -8,11 +8,11 @@ import (
"sync"
"time"
cclog "github.com/ClusterCockpit/cc-lib/ccLogger"
cclog "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
lp "github.com/ClusterCockpit/cc-lib/ccMessage"
mp "github.com/ClusterCockpit/cc-lib/messageProcessor"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
agg "github.com/ClusterCockpit/cc-metric-collector/internal/metricAggregator"
mp "github.com/ClusterCockpit/cc-metric-collector/pkg/messageProcessor"
mct "github.com/ClusterCockpit/cc-metric-collector/pkg/multiChanTicker"
)
@@ -63,7 +63,7 @@ type metricRouter struct {
// MetricRouter access functions
type MetricRouter interface {
Init(ticker mct.MultiChanTicker, wg *sync.WaitGroup, routerConfig json.RawMessage) error
Init(ticker mct.MultiChanTicker, wg *sync.WaitGroup, routerConfigFile string) error
AddCollectorInput(input chan lp.CCMessage)
AddReceiverInput(input chan lp.CCMessage)
AddOutput(output chan lp.CCMessage)
@@ -77,7 +77,7 @@ type MetricRouter interface {
// * wait group synchronization (from variable wg)
// * ticker (from variable ticker)
// * configuration (read from config file in variable routerConfigFile)
func (r *metricRouter) Init(ticker mct.MultiChanTicker, wg *sync.WaitGroup, routerConfig json.RawMessage) error {
func (r *metricRouter) Init(ticker mct.MultiChanTicker, wg *sync.WaitGroup, routerConfigFile string) error {
r.outputs = make([]chan lp.CCMessage, 0)
r.done = make(chan bool)
r.cache_input = make(chan lp.CCMessage)
@@ -95,7 +95,15 @@ func (r *metricRouter) Init(ticker mct.MultiChanTicker, wg *sync.WaitGroup, rout
// Drop domain part of host name
r.hostname = strings.SplitN(hostname, `.`, 2)[0]
err = json.Unmarshal(routerConfig, &r.config)
// Read metric router config file
configFile, err := os.Open(routerConfigFile)
if err != nil {
cclog.ComponentError("MetricRouter", err.Error())
return err
}
defer configFile.Close()
jsonParser := json.NewDecoder(configFile)
err = jsonParser.Decode(&r.config)
if err != nil {
cclog.ComponentError("MetricRouter", err.Error())
return err
@@ -441,9 +449,9 @@ func (r *metricRouter) Close() {
}
// New creates a new initialized metric router
func New(ticker mct.MultiChanTicker, wg *sync.WaitGroup, routerConfig json.RawMessage) (MetricRouter, error) {
func New(ticker mct.MultiChanTicker, wg *sync.WaitGroup, routerConfigFile string) (MetricRouter, error) {
r := new(metricRouter)
err := r.Init(ticker, wg, routerConfig)
err := r.Init(ticker, wg, routerConfigFile)
if err != nil {
return nil, err
}

113
pkg/ccLogger/cclogger.go Normal file
View File

@@ -0,0 +1,113 @@
package cclogger
import (
"fmt"
"log"
"os"
"runtime"
)
var (
globalDebug = false
stdout = os.Stdout
stderr = os.Stderr
debugLog *log.Logger = nil
infoLog *log.Logger = nil
errorLog *log.Logger = nil
warnLog *log.Logger = nil
defaultLog *log.Logger = nil
)
func initLogger() {
if debugLog == nil {
debugLog = log.New(stderr, "DEBUG ", log.LstdFlags)
}
if infoLog == nil {
infoLog = log.New(stdout, "INFO ", log.LstdFlags)
}
if errorLog == nil {
errorLog = log.New(stderr, "ERROR ", log.LstdFlags)
}
if warnLog == nil {
warnLog = log.New(stderr, "WARN ", log.LstdFlags)
}
if defaultLog == nil {
defaultLog = log.New(stdout, "", log.LstdFlags)
}
}
func Print(e ...interface{}) {
initLogger()
defaultLog.Print(e...)
}
func ComponentPrint(component string, e ...interface{}) {
initLogger()
defaultLog.Print(fmt.Sprintf("[%s] ", component), e)
}
func Info(e ...interface{}) {
initLogger()
infoLog.Print(e...)
}
func ComponentInfo(component string, e ...interface{}) {
initLogger()
infoLog.Print(fmt.Sprintf("[%s] ", component), e)
}
func Debug(e ...interface{}) {
initLogger()
if globalDebug {
debugLog.Print(e...)
}
}
func ComponentDebug(component string, e ...interface{}) {
initLogger()
if globalDebug && debugLog != nil {
//CCComponentPrint(debugLog, component, e)
debugLog.Print(fmt.Sprintf("[%s] ", component), e)
}
}
func Error(e ...interface{}) {
initLogger()
_, fn, line, _ := runtime.Caller(1)
errorLog.Print(fmt.Sprintf("[%s:%d] ", fn, line), e)
}
func ComponentError(component string, e ...interface{}) {
initLogger()
_, fn, line, _ := runtime.Caller(1)
errorLog.Print(fmt.Sprintf("[%s|%s:%d] ", component, fn, line), e)
}
func SetDebug() {
globalDebug = true
initLogger()
}
func SetOutput(filename string) {
if filename == "stderr" {
if stderr != os.Stderr && stderr != os.Stdout {
stderr.Close()
}
stderr = os.Stderr
} else if filename == "stdout" {
if stderr != os.Stderr && stderr != os.Stdout {
stderr.Close()
}
stderr = os.Stdout
} else {
file, err := os.OpenFile(filename, os.O_APPEND|os.O_CREATE|os.O_WRONLY, 0600)
if err == nil {
defer file.Close()
stderr = file
}
}
debugLog = nil
errorLog = nil
warnLog = nil
initLogger()
}

57
pkg/ccMetric/README.md Normal file
View File

@@ -0,0 +1,57 @@
# ClusterCockpit metrics
As described in the [ClusterCockpit specifications](https://github.com/ClusterCockpit/cc-specifications), the whole ClusterCockpit stack uses metrics in the InfluxDB line protocol format. This is also the input and output format for the ClusterCockpit Metric Collector but internally it uses an extended format while processing, named CCMetric.
It is basically a copy of the [InfluxDB line protocol](https://github.com/influxdata/line-protocol) `MutableMetric` interface with one extension. Besides the tags and fields, it contains a list of meta information (re-using the `Tag` structure of the original protocol):
```golang
type ccMetric struct {
name string // Measurement name
meta map[string]string // map of meta data tags
tags map[string]string // map of of tags
fields map[string]interface{} // map of of fields
tm time.Time // timestamp
}
type CCMetric interface {
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
String() string // Return line-protocol like string
Name() string // Get metric name
SetName(name string) // Set metric name
Time() time.Time // Get timestamp
SetTime(t time.Time) // Set timestamp
Tags() map[string]string // Map of tags
AddTag(key, value string) // Add a tag
GetTag(key string) (value string, ok bool) // Get a tag by its key
HasTag(key string) (ok bool) // Check if a tag key is present
RemoveTag(key string) // Remove a tag by its key
Meta() map[string]string // Map of meta data tags
AddMeta(key, value string) // Add a meta data tag
GetMeta(key string) (value string, ok bool) // Get a meta data tab addressed by its key
HasMeta(key string) (ok bool) // Check if a meta data key is present
RemoveMeta(key string) // Remove a meta data tag by its key
Fields() map[string]interface{} // Map of fields
AddField(key string, value interface{}) // Add a field
GetField(key string) (value interface{}, ok bool) // Get a field addressed by its key
HasField(key string) (ok bool) // Check if a field key is present
RemoveField(key string) // Remove a field addressed by its key
}
func New(name string, tags map[string]string, meta map[string]string, fields map[string]interface{}, tm time.Time) (CCMetric, error)
func FromMetric(other CCMetric) CCMetric
func FromInfluxMetric(other lp.Metric) CCMetric
```
The `CCMetric` interface provides the same functions as the `MutableMetric` like `{Add, Get, Remove, Has}{Tag, Field}` and additionally provides `{Add, Get, Remove, Has}Meta`.
The InfluxDB protocol creates a new metric with `influx.New(name, tags, fields, time)` while CCMetric uses `ccMetric.New(name, tags, meta, fields, time)` where `tags` and `meta` are both of type `map[string]string`.
You can copy a CCMetric with `FromMetric(other CCMetric) CCMetric`. If you get an `influx.Metric` from a function, like the line protocol parser, you can use `FromInfluxMetric(other influx.Metric) CCMetric` to get a CCMetric out of it (see `NatsReceiver` for an example).
Although the [cc-specifications](https://github.com/ClusterCockpit/cc-specifications/blob/master/interfaces/lineprotocol/README.md) defines that there is only a `value` field for the metric value, the CCMetric still can have multiple values similar to the InfluxDB line protocol.

353
pkg/ccMetric/ccMetric.go Normal file
View File

@@ -0,0 +1,353 @@
package ccmetric
import (
"fmt"
"time"
influxdb2 "github.com/influxdata/influxdb-client-go/v2"
write "github.com/influxdata/influxdb-client-go/v2/api/write"
lp "github.com/influxdata/line-protocol" // MIT license
"golang.org/x/exp/maps"
)
// Most functions are derived from github.com/influxdata/line-protocol/metric.go
// The metric type is extended with an extra meta information list re-using the Tag
// type.
//
// See: https://docs.influxdata.com/influxdb/latest/reference/syntax/line-protocol/
type ccMetric struct {
name string // Measurement name
meta map[string]string // map of meta data tags
tags map[string]string // map of of tags
fields map[string]interface{} // map of of fields
tm time.Time // timestamp
}
// ccMetric access functions
type CCMetric interface {
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
Time() time.Time // Get timestamp
SetTime(t time.Time) // Set timestamp
Tags() map[string]string // Map of tags
AddTag(key, value string) // Add a tag
GetTag(key string) (value string, ok bool) // Get a tag by its key
HasTag(key string) (ok bool) // Check if a tag key is present
RemoveTag(key string) // Remove a tag by its key
Meta() map[string]string // Map of meta data tags
AddMeta(key, value string) // Add a meta data tag
GetMeta(key string) (value string, ok bool) // Get a meta data tab addressed by its key
HasMeta(key string) (ok bool) // Check if a meta data key is present
RemoveMeta(key string) // Remove a meta data tag by its key
Fields() map[string]interface{} // Map of fields
AddField(key string, value interface{}) // Add a field
GetField(key string) (value interface{}, ok bool) // Get a field addressed by its key
HasField(key string) (ok bool) // Check if a field key is present
RemoveField(key string) // Remove a field addressed by its key
String() string // Return line-protocol like string
}
// String implements the stringer interface for data type ccMetric
func (m *ccMetric) String() string {
return fmt.Sprintf(
"Name: %s, Tags: %+v, Meta: %+v, fields: %+v, Timestamp: %d",
m.name, m.tags, m.meta, m.fields, m.tm.UnixNano(),
)
}
// ToLineProtocol generates influxDB line protocol for data type ccMetric
func (m *ccMetric) ToPoint(metaAsTags map[string]bool) (p *write.Point) {
p = influxdb2.NewPoint(m.name, m.tags, m.fields, m.tm)
for key, use_as_tag := range metaAsTags {
if use_as_tag {
if value, ok := m.GetMeta(key); ok {
p.AddTag(key, value)
}
}
}
return p
}
// ToLineProtocol generates influxDB line protocol for data type ccMetric
func (m *ccMetric) ToLineProtocol(metaAsTags map[string]bool) string {
return write.PointToLineProtocol(
m.ToPoint(metaAsTags),
time.Nanosecond,
)
}
// Name returns the measurement name
func (m *ccMetric) Name() string {
return m.name
}
// SetName sets the measurement name
func (m *ccMetric) SetName(name string) {
m.name = name
}
// Time returns timestamp
func (m *ccMetric) Time() time.Time {
return m.tm
}
// SetTime sets the timestamp
func (m *ccMetric) SetTime(t time.Time) {
m.tm = t
}
// Tags returns the the list of tags as key-value-mapping
func (m *ccMetric) Tags() map[string]string {
return m.tags
}
// AddTag adds a tag (consisting of key and value) to the map of tags
func (m *ccMetric) AddTag(key, value string) {
m.tags[key] = value
}
// GetTag returns the tag with tag's key equal to <key>
func (m *ccMetric) GetTag(key string) (string, bool) {
value, ok := m.tags[key]
return value, ok
}
// HasTag checks if a tag with key equal to <key> is present in the list of tags
func (m *ccMetric) HasTag(key string) bool {
_, ok := m.tags[key]
return ok
}
// RemoveTag removes the tag with tag's key equal to <key>
func (m *ccMetric) RemoveTag(key string) {
delete(m.tags, key)
}
// Meta returns the meta data tags as key-value mapping
func (m *ccMetric) Meta() map[string]string {
return m.meta
}
// AddMeta adds a meta data tag (consisting of key and value) to the map of meta data tags
func (m *ccMetric) AddMeta(key, value string) {
m.meta[key] = value
}
// GetMeta returns the meta data tag with meta data's key equal to <key>
func (m *ccMetric) GetMeta(key string) (string, bool) {
value, ok := m.meta[key]
return value, ok
}
// HasMeta checks if a meta data tag with meta data's key equal to <key> is present in the map of meta data tags
func (m *ccMetric) HasMeta(key string) bool {
_, ok := m.meta[key]
return ok
}
// RemoveMeta removes the meta data tag with tag's key equal to <key>
func (m *ccMetric) RemoveMeta(key string) {
delete(m.meta, key)
}
// Fields returns the list of fields as key-value-mapping
func (m *ccMetric) Fields() map[string]interface{} {
return m.fields
}
// AddField adds a field (consisting of key and value) to the map of fields
func (m *ccMetric) AddField(key string, value interface{}) {
m.fields[key] = value
}
// GetField returns the field with field's key equal to <key>
func (m *ccMetric) GetField(key string) (interface{}, bool) {
v, ok := m.fields[key]
return v, ok
}
// HasField checks if a field with field's key equal to <key> is present in the map of fields
func (m *ccMetric) HasField(key string) bool {
_, ok := m.fields[key]
return ok
}
// RemoveField removes the field with field's key equal to <key>
// from the map of fields
func (m *ccMetric) RemoveField(key string) {
delete(m.fields, key)
}
// New creates a new measurement point
func New(
name string,
tags map[string]string,
meta map[string]string,
fields map[string]interface{},
tm time.Time,
) (CCMetric, error) {
m := &ccMetric{
name: name,
tags: maps.Clone(tags),
meta: maps.Clone(meta),
fields: make(map[string]interface{}, len(fields)),
tm: tm,
}
// deep copy fields
for k, v := range fields {
v := convertField(v)
if v == nil {
continue
}
m.fields[k] = v
}
return m, nil
}
// FromMetric copies the metric <other>
func FromMetric(other CCMetric) CCMetric {
return &ccMetric{
name: other.Name(),
tags: maps.Clone(other.Tags()),
meta: maps.Clone(other.Meta()),
fields: maps.Clone(other.Fields()),
tm: other.Time(),
}
}
// FromInfluxMetric copies the influxDB line protocol metric <other>
func FromInfluxMetric(other lp.Metric) CCMetric {
m := &ccMetric{
name: other.Name(),
tags: make(map[string]string),
meta: make(map[string]string),
fields: make(map[string]interface{}),
tm: other.Time(),
}
// deep copy tags and fields
for _, otherTag := range other.TagList() {
m.tags[otherTag.Key] = otherTag.Value
}
for _, otherField := range other.FieldList() {
m.fields[otherField.Key] = otherField.Value
}
return m
}
// convertField converts data types of fields by the following schemata:
//
// *float32, *float64, float32, float64 -> float64
// *int, *int8, *int16, *int32, *int64, int, int8, int16, int32, int64 -> int64
//
// *uint, *uint8, *uint16, *uint32, *uint64, uint, uint8, uint16, uint32, uint64 -> uint64
// *[]byte, *string, []byte, string -> string
// *bool, bool -> bool
func convertField(v interface{}) interface{} {
switch v := v.(type) {
case float64:
return v
case int64:
return v
case string:
return v
case bool:
return v
case int:
return int64(v)
case uint:
return uint64(v)
case uint64:
return uint64(v)
case []byte:
return string(v)
case int32:
return int64(v)
case int16:
return int64(v)
case int8:
return int64(v)
case uint32:
return uint64(v)
case uint16:
return uint64(v)
case uint8:
return uint64(v)
case float32:
return float64(v)
case *float64:
if v != nil {
return *v
}
case *int64:
if v != nil {
return *v
}
case *string:
if v != nil {
return *v
}
case *bool:
if v != nil {
return *v
}
case *int:
if v != nil {
return int64(*v)
}
case *uint:
if v != nil {
return uint64(*v)
}
case *uint64:
if v != nil {
return uint64(*v)
}
case *[]byte:
if v != nil {
return string(*v)
}
case *int32:
if v != nil {
return int64(*v)
}
case *int16:
if v != nil {
return int64(*v)
}
case *int8:
if v != nil {
return int64(*v)
}
case *uint32:
if v != nil {
return uint64(*v)
}
case *uint16:
if v != nil {
return uint64(*v)
}
case *uint8:
if v != nil {
return uint64(*v)
}
case *float32:
if v != nil {
return float64(*v)
}
default:
return nil
}
return nil
}

View File

@@ -9,7 +9,7 @@ import (
"strconv"
"strings"
cclogger "github.com/ClusterCockpit/cc-lib/ccLogger"
cclogger "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
"golang.org/x/exp/slices"
)

View File

@@ -0,0 +1,266 @@
# Message Processor Component
Multiple parts of in the ClusterCockit ecosystem require the processing of CCMessages.
The main CC application using it is `cc-metric-collector`. The processing part there was originally in the metric router, the central
hub connecting collectors (reading local data), receivers (receiving remote data) and sinks (sending data). Already in early stages, the
lack of flexibility caused some trouble:
> The sysadmins wanted to keep operating their Ganglia based monitoring infrastructure while we developed the CC stack. Ganglia wants the core metrics with
> a specific name and resolution (right unit prefix) but there was no conversion of the data in the CC stack, so CC frontend developers wanted a different
> resolution for some metrics. The issue was basically the `mem_used` metric showing the currently used memory of the node. Ganglia wants it in `kByte` as provided
> by the Linux operating system but CC wanted it in `GByte`.
With the message processor, the Ganglia sinks can apply the unit prefix changes individually and name the metrics as required by Ganglia.
## For developers
Whenever you receive or are about to send a message out, you should provide some processing.
### Configuration of component
New operations can be added to the message processor at runtime. Of course, they can also be removed again. For the initial setup, having a configuration file
or some fields in a configuration file for the processing.
The message processor uses the following configuration
```json
{
"drop_messages": [
"name_of_message_to_drop"
],
"drop_messages_if": [
"condition_when_to_drop_message",
"name == 'drop_this'",
"tag.hostname == 'this_host'",
"meta.unit != 'MB'"
],
"rename_messages" : {
"old_message_name" : "new_message_name"
},
"rename_messages_if": {
"condition_when_to_rename_message" : "new_name"
},
"add_tags_if": [
{
"if" : "condition_when_to_add_tag",
"key": "name_for_new_tag",
"value": "new_tag_value"
}
],
"delete_tags_if": [
{
"if" : "condition_when_to_delete_tag",
"key": "name_of_tag"
}
],
"add_meta_if": [
{
"if" : "condition_when_to_add_meta_info",
"key": "name_for_new_meta_info",
"value": "new_meta_info_value"
}
],
"delete_meta_if": [
{
"if" : "condition_when_to_delete_meta_info",
"key": "name_of_meta_info"
}
],
"add_field_if": [
{
"if" : "condition_when_to_add_field",
"key": "name_for_new_field",
"value": "new_field_value_but_only_string_at_the_moment"
}
],
"delete_field_if": [
{
"if" : "condition_when_to_delete_field",
"key": "name_of_field"
}
],
"move_tag_to_meta_if": [
{
"if" : "condition_when_to_move_tag_to_meta_info_including_its_value",
"key": "name_of_tag",
"value": "name_of_meta_info"
}
],
"move_tag_to_field_if": [
{
"if" : "condition_when_to_move_tag_to_fields_including_its_value",
"key": "name_of_tag",
"value": "name_of_field"
}
],
"move_meta_to_tag_if": [
{
"if" : "condition_when_to_move_meta_info_to_tags_including_its_value",
"key": "name_of_meta_info",
"value": "name_of_tag"
}
],
"move_meta_to_field_if": [
{
"if" : "condition_when_to_move_meta_info_to_fields_including_its_value",
"key": "name_of_tag",
"value": "name_of_meta_info"
}
],
"move_field_to_tag_if": [
{
"if" : "condition_when_to_move_field_to_tags_including_its_stringified_value",
"key": "name_of_field",
"value": "name_of_tag"
}
],
"move_field_to_meta_if": [
{
"if" : "condition_when_to_move_field_to_meta_info_including_its_stringified_value",
"key": "name_of_field",
"value": "name_of_meta_info"
}
],
"drop_by_message_type": [
"metric",
"event",
"log",
"control"
],
"change_unit_prefix": {
"name == 'metric_with_wrong_unit_prefix'" : "G",
"only_if_messagetype == 'metric'": "T"
},
"normalize_units": true,
"add_base_env": {
"MY_CONSTANT_FOR_CUSTOM_CONDITIONS": 1.0,
"output_value_for_test_metrics": 42.0,
},
"stage_order": [
"rename_messages_if",
"drop_messages"
]
}
```
The options `change_unit_prefix` and `normalize_units` are only applied to CCMetrics. It is not possible to delete the field related to each message type as defined in [cc-specification](https://github.com/ClusterCockpit/cc-specifications/tree/master/interfaces/lineprotocol). In short:
- CCMetrics always have to have a field named `value`
- CCEvents always have to have a field named `event`
- CCLogs always have to have a field named `log`
- CCControl messages always have to have a field named `control`
With `add_base_env`, one can specifiy mykey=myvalue pairs that can be used in conditions like `tag.type == mykey`.
The order in which each message is processed, can be specified with the `stage_order` option. The stage names are the keys in the JSON configuration, thus `change_unit_prefix`, `move_field_to_meta_if`, etc. Stages can be listed multiple times.
### Using the component
In order to load the configuration from a `json.RawMessage`:
```golang
mp, err := NewMessageProcessor()
if err != nil {
log.Error("failed to create new message processor")
}
mp.FromConfigJSON(configJson)
```
After initialization and adding the different operations, the `ProcessMessage()` function applies all operations and returns whether the message should be dropped.
```golang
m := lp.CCMetric{}
x, err := mp.ProcessMessage(m)
if err != nil {
// handle error
}
if x != nil {
// process x further
} else {
// this message got dropped
}
```
Single operations can be added and removed at runtime
```golang
type MessageProcessor interface {
// Functions to set the execution order of the processing stages
SetStages([]string) error
DefaultStages() []string
// Function to add variables to the base evaluation environment
AddBaseEnv(env map[string]interface{}) error
// Functions to add and remove rules
AddDropMessagesByName(name string) error
RemoveDropMessagesByName(name string)
AddDropMessagesByCondition(condition string) error
RemoveDropMessagesByCondition(condition string)
AddRenameMetricByCondition(condition string, name string) error
RemoveRenameMetricByCondition(condition string)
AddRenameMetricByName(from, to string) error
RemoveRenameMetricByName(from string)
SetNormalizeUnits(settings bool)
AddChangeUnitPrefix(condition string, prefix string) error
RemoveChangeUnitPrefix(condition string)
AddAddTagsByCondition(condition, key, value string) error
RemoveAddTagsByCondition(condition string)
AddDeleteTagsByCondition(condition, key, value string) error
RemoveDeleteTagsByCondition(condition string)
AddAddMetaByCondition(condition, key, value string) error
RemoveAddMetaByCondition(condition string)
AddDeleteMetaByCondition(condition, key, value string) error
RemoveDeleteMetaByCondition(condition string)
AddMoveTagToMeta(condition, key, value string) error
RemoveMoveTagToMeta(condition string)
AddMoveTagToFields(condition, key, value string) error
RemoveMoveTagToFields(condition string)
AddMoveMetaToTags(condition, key, value string) error
RemoveMoveMetaToTags(condition string)
AddMoveMetaToFields(condition, key, value string) error
RemoveMoveMetaToFields(condition string)
AddMoveFieldToTags(condition, key, value string) error
RemoveMoveFieldToTags(condition string)
AddMoveFieldToMeta(condition, key, value string) error
RemoveMoveFieldToMeta(condition string)
// Read in a JSON configuration
FromConfigJSON(config json.RawMessage) error
ProcessMessage(m lp2.CCMessage) (lp2.CCMessage, error)
// Processing functions for legacy CCMetric and current CCMessage
ProcessMetric(m lp.CCMetric) (lp2.CCMessage, error)
}
```
### Syntax for evaluatable terms
The message processor uses `gval` for evaluating the terms. It provides a basic set of operators like string comparison and arithmetic operations.
Accessible for operations are
- `name` of the message
- `timestamp` or `time` of the message
- `type`, `type-id` of the message (also `tag_type`, `tag_type-id` and `tag_typeid`)
- `stype`, `stype-id` of the message (if message has theses tags, also `tag_stype`, `tag_stype-id` and `tag_stypeid`)
- `value` for a CCMetric message (also `field_value`)
- `event` for a CCEvent message (also `field_event`)
- `control` for a CCControl message (also `field_control`)
- `log` for a CCLog message (also `field_log`)
- `messagetype` or `msgtype`. Possible values `event`, `metric`, `log` and `control`.
Generally, all tags are accessible with `tag_<tagkey>`, `tags_<tagkey>` or `tags.<tagkey>`. Similarly for all fields with `field[s]?[_.]<fieldkey>`. For meta information `meta[_.]<metakey>` (there is no `metas[_.]<metakey>`).
The [syntax of `expr`](https://expr-lang.org/docs/language-definition) is accepted with some additions:
- Comparing strings: `==`, `!=`, `str matches regex` (use `%` instead of `\`!)
- Combining conditions: `&&`, `||`
- Comparing numbers: `==`, `!=`, `<`, `>`, `<=`, `>=`
- Test lists: `<value> in <list>`
- Topological tests: `tag_type-id in getCpuListOfType("socket", "1")` (test if the metric belongs to socket 1 in local node topology)
Often the operations are written in JSON files for loading them at startup. In JSON, some characters are not allowed. Therefore, the term syntax reflects that:
- use `''` instead of `""` for strings
- for the regexes, use `%` instead of `\`
For operations that should be applied on all messages, use the condition `true`.
### Overhead
The operations taking conditions are pre-processed, which is commonly the time consuming part but, of course, with each added operation, the time to process a message
increases. Moreover, the processing creates a copy of the message.

View File

@@ -0,0 +1,988 @@
package messageprocessor
import (
"encoding/json"
"fmt"
"strings"
"sync"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
cclog "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
lplegacy "github.com/ClusterCockpit/cc-metric-collector/pkg/ccMetric"
"github.com/expr-lang/expr"
"github.com/expr-lang/expr/vm"
)
// Message processor add/delete tag/meta configuration
type messageProcessorTagConfig struct {
Key string `json:"key"` // Tag name
Value string `json:"value,omitempty"` // Tag value
Condition string `json:"if"` // Condition for adding or removing corresponding tag
}
type messageProcessorConfig struct {
StageOrder []string `json:"stage_order,omitempty"` // List of stages to execute them in the specified order and to skip unrequired ones
DropMessages []string `json:"drop_messages,omitempty"` // List of metric names to drop. For fine-grained dropping use drop_messages_if
DropMessagesIf []string `json:"drop_messages_if,omitempty"` // List of evaluatable terms to drop messages
RenameMessages map[string]string `json:"rename_messages,omitempty"` // Map of metric names to rename
RenameMessagesIf map[string]string `json:"rename_messages_if,omitempty"` // Map to rename metric name based on a condition
NormalizeUnits bool `json:"normalize_units,omitempty"` // Check unit meta flag and normalize it using cc-units
ChangeUnitPrefix map[string]string `json:"change_unit_prefix,omitempty"` // Add prefix that should be applied to the messages
AddTagsIf []messageProcessorTagConfig `json:"add_tags_if,omitempty"` // List of tags that are added when the condition is met
DelTagsIf []messageProcessorTagConfig `json:"delete_tags_if,omitempty"` // List of tags that are removed when the condition is met
AddMetaIf []messageProcessorTagConfig `json:"add_meta_if,omitempty"` // List of meta infos that are added when the condition is met
DelMetaIf []messageProcessorTagConfig `json:"delete_meta_if,omitempty"` // List of meta infos that are removed when the condition is met
AddFieldIf []messageProcessorTagConfig `json:"add_field_if,omitempty"` // List of fields that are added when the condition is met
DelFieldIf []messageProcessorTagConfig `json:"delete_field_if,omitempty"` // List of fields that are removed when the condition is met
DropByType []string `json:"drop_by_message_type,omitempty"` // List of message types that should be dropped
MoveTagToMeta []messageProcessorTagConfig `json:"move_tag_to_meta_if,omitempty"`
MoveTagToField []messageProcessorTagConfig `json:"move_tag_to_field_if,omitempty"`
MoveMetaToTag []messageProcessorTagConfig `json:"move_meta_to_tag_if,omitempty"`
MoveMetaToField []messageProcessorTagConfig `json:"move_meta_to_field_if,omitempty"`
MoveFieldToTag []messageProcessorTagConfig `json:"move_field_to_tag_if,omitempty"`
MoveFieldToMeta []messageProcessorTagConfig `json:"move_field_to_meta_if,omitempty"`
AddBaseEnv map[string]interface{} `json:"add_base_env,omitempty"`
}
type messageProcessor struct {
// For thread-safety
mutex sync.RWMutex
// mapping contains all evalables as strings to gval.Evaluable
// because it is not possible to get the original string out of
// a gval.Evaluable
mapping map[string]*vm.Program
stages []string // order of stage execution
dropMessages map[string]struct{} // internal lookup map
dropTypes map[string]struct{} // internal lookup map
dropMessagesIf map[*vm.Program]struct{} // pre-processed dropMessagesIf
renameMessages map[string]string // internal lookup map
renameMessagesIf map[*vm.Program]string // pre-processed RenameMessagesIf
changeUnitPrefix map[*vm.Program]string // pre-processed ChangeUnitPrefix
normalizeUnits bool
addTagsIf map[*vm.Program]messageProcessorTagConfig // pre-processed AddTagsIf
deleteTagsIf map[*vm.Program]messageProcessorTagConfig // pre-processed DelTagsIf
addMetaIf map[*vm.Program]messageProcessorTagConfig // pre-processed AddMetaIf
deleteMetaIf map[*vm.Program]messageProcessorTagConfig // pre-processed DelMetaIf
addFieldIf map[*vm.Program]messageProcessorTagConfig // pre-processed AddFieldIf
deleteFieldIf map[*vm.Program]messageProcessorTagConfig // pre-processed DelFieldIf
moveTagToMeta map[*vm.Program]messageProcessorTagConfig // pre-processed MoveTagToMeta
moveTagToField map[*vm.Program]messageProcessorTagConfig // pre-processed MoveTagToField
moveMetaToTag map[*vm.Program]messageProcessorTagConfig // pre-processed MoveMetaToTag
moveMetaToField map[*vm.Program]messageProcessorTagConfig // pre-processed MoveMetaToField
moveFieldToTag map[*vm.Program]messageProcessorTagConfig // pre-processed MoveFieldToTag
moveFieldToMeta map[*vm.Program]messageProcessorTagConfig // pre-processed MoveFieldToMeta
}
type MessageProcessor interface {
// Functions to set the execution order of the processing stages
SetStages([]string) error
DefaultStages() []string
// Function to add variables to the base evaluation environment
AddBaseEnv(env map[string]interface{}) error
// Functions to add and remove rules
AddDropMessagesByName(name string) error
RemoveDropMessagesByName(name string)
AddDropMessagesByCondition(condition string) error
RemoveDropMessagesByCondition(condition string)
AddRenameMetricByCondition(condition string, name string) error
RemoveRenameMetricByCondition(condition string)
AddRenameMetricByName(from, to string) error
RemoveRenameMetricByName(from string)
SetNormalizeUnits(settings bool)
AddChangeUnitPrefix(condition string, prefix string) error
RemoveChangeUnitPrefix(condition string)
AddAddTagsByCondition(condition, key, value string) error
RemoveAddTagsByCondition(condition string)
AddDeleteTagsByCondition(condition, key, value string) error
RemoveDeleteTagsByCondition(condition string)
AddAddMetaByCondition(condition, key, value string) error
RemoveAddMetaByCondition(condition string)
AddDeleteMetaByCondition(condition, key, value string) error
RemoveDeleteMetaByCondition(condition string)
AddMoveTagToMeta(condition, key, value string) error
RemoveMoveTagToMeta(condition string)
AddMoveTagToFields(condition, key, value string) error
RemoveMoveTagToFields(condition string)
AddMoveMetaToTags(condition, key, value string) error
RemoveMoveMetaToTags(condition string)
AddMoveMetaToFields(condition, key, value string) error
RemoveMoveMetaToFields(condition string)
AddMoveFieldToTags(condition, key, value string) error
RemoveMoveFieldToTags(condition string)
AddMoveFieldToMeta(condition, key, value string) error
RemoveMoveFieldToMeta(condition string)
// Read in a JSON configuration
FromConfigJSON(config json.RawMessage) error
// Processing functions for legacy CCMetric and current CCMessage
ProcessMetric(m lplegacy.CCMetric) (lp.CCMessage, error)
ProcessMessage(m lp.CCMessage) (lp.CCMessage, error)
//EvalToBool(condition string, parameters map[string]interface{}) (bool, error)
//EvalToFloat64(condition string, parameters map[string]interface{}) (float64, error)
//EvalToString(condition string, parameters map[string]interface{}) (string, error)
}
const (
STAGENAME_DROP_BY_NAME string = "drop_by_name"
STAGENAME_DROP_BY_TYPE string = "drop_by_type"
STAGENAME_DROP_IF string = "drop_if"
STAGENAME_ADD_TAG string = "add_tag"
STAGENAME_DELETE_TAG string = "delete_tag"
STAGENAME_MOVE_TAG_META string = "move_tag_to_meta"
STAGENAME_MOVE_TAG_FIELD string = "move_tag_to_fields"
STAGENAME_ADD_META string = "add_meta"
STAGENAME_DELETE_META string = "delete_meta"
STAGENAME_MOVE_META_TAG string = "move_meta_to_tags"
STAGENAME_MOVE_META_FIELD string = "move_meta_to_fields"
STAGENAME_ADD_FIELD string = "add_field"
STAGENAME_DELETE_FIELD string = "delete_field"
STAGENAME_MOVE_FIELD_TAG string = "move_field_to_tags"
STAGENAME_MOVE_FIELD_META string = "move_field_to_meta"
STAGENAME_RENAME_BY_NAME string = "rename"
STAGENAME_RENAME_IF string = "rename_if"
STAGENAME_CHANGE_UNIT_PREFIX string = "change_unit_prefix"
STAGENAME_NORMALIZE_UNIT string = "normalize_unit"
)
var StageNames = []string{
STAGENAME_DROP_BY_NAME,
STAGENAME_DROP_BY_TYPE,
STAGENAME_DROP_IF,
STAGENAME_ADD_TAG,
STAGENAME_DELETE_TAG,
STAGENAME_MOVE_TAG_META,
STAGENAME_MOVE_TAG_FIELD,
STAGENAME_ADD_META,
STAGENAME_DELETE_META,
STAGENAME_MOVE_META_TAG,
STAGENAME_MOVE_META_FIELD,
STAGENAME_ADD_FIELD,
STAGENAME_DELETE_FIELD,
STAGENAME_MOVE_FIELD_TAG,
STAGENAME_MOVE_FIELD_META,
STAGENAME_RENAME_BY_NAME,
STAGENAME_RENAME_IF,
STAGENAME_CHANGE_UNIT_PREFIX,
STAGENAME_NORMALIZE_UNIT,
}
var paramMapPool = sync.Pool{
New: func() any {
return make(map[string]interface{})
},
}
func sanitizeExprString(key string) string {
return strings.ReplaceAll(key, "type-id", "typeid")
}
func getParamMap(point lp.CCMetric) map[string]interface{} {
params := paramMapPool.Get().(map[string]interface{})
params["message"] = point
params["msg"] = point
params["name"] = point.Name()
params["timestamp"] = point.Time().Unix()
params["time"] = params["timestamp"]
fields := paramMapPool.Get().(map[string]interface{})
for key, value := range point.Fields() {
fields[key] = value
switch key {
case "value":
params["messagetype"] = "metric"
params["value"] = value
params["metric"] = value
case "event":
params["messagetype"] = "event"
params["event"] = value
case "control":
params["messagetype"] = "control"
params["control"] = value
case "log":
params["messagetype"] = "log"
params["log"] = value
default:
params["messagetype"] = "unknown"
}
}
params["msgtype"] = params["messagetype"]
params["fields"] = fields
params["field"] = fields
tags := paramMapPool.Get().(map[string]interface{})
for key, value := range point.Tags() {
tags[sanitizeExprString(key)] = value
}
params["tags"] = tags
params["tag"] = tags
meta := paramMapPool.Get().(map[string]interface{})
for key, value := range point.Meta() {
meta[sanitizeExprString(key)] = value
}
params["meta"] = meta
return params
}
var baseenv = map[string]interface{}{
"name": "",
"messagetype": "unknown",
"msgtype": "unknown",
"tag": map[string]interface{}{
"type": "unknown",
"typeid": "0",
"stype": "unknown",
"stypeid": "0",
"hostname": "localhost",
"cluster": "nocluster",
},
"tags": map[string]interface{}{
"type": "unknown",
"typeid": "0",
"stype": "unknown",
"stypeid": "0",
"hostname": "localhost",
"cluster": "nocluster",
},
"meta": map[string]interface{}{
"unit": "invalid",
"source": "unknown",
},
"fields": map[string]interface{}{
"value": 0,
"event": "",
"control": "",
"log": "",
},
"field": map[string]interface{}{
"value": 0,
"event": "",
"control": "",
"log": "",
},
"timestamp": 1234567890,
"msg": lp.EmptyMessage(),
"message": lp.EmptyMessage(),
}
func addBaseEnvWalker(values map[string]interface{}) map[string]interface{} {
out := make(map[string]interface{})
for k, v := range values {
switch value := v.(type) {
case int, int32, int64, uint, uint32, uint64, string, float32, float64:
out[k] = value
case map[string]interface{}:
if _, ok := baseenv[k]; !ok {
out[k] = addBaseEnvWalker(value)
}
}
}
return out
}
func (mp *messageProcessor) AddBaseEnv(env map[string]interface{}) error {
for k, v := range env {
switch value := v.(type) {
case int, int32, int64, uint, uint32, uint64, string, float32, float64:
baseenv[k] = value
case map[string]interface{}:
if _, ok := baseenv[k]; !ok {
baseenv[k] = addBaseEnvWalker(value)
}
}
}
return nil
}
func (mp *messageProcessor) init() error {
mp.stages = make([]string, 0)
mp.mapping = make(map[string]*vm.Program)
mp.dropMessages = make(map[string]struct{})
mp.dropTypes = make(map[string]struct{})
mp.dropMessagesIf = make(map[*vm.Program]struct{})
mp.renameMessages = make(map[string]string)
mp.renameMessagesIf = make(map[*vm.Program]string)
mp.changeUnitPrefix = make(map[*vm.Program]string)
mp.addTagsIf = make(map[*vm.Program]messageProcessorTagConfig)
mp.addMetaIf = make(map[*vm.Program]messageProcessorTagConfig)
mp.addFieldIf = make(map[*vm.Program]messageProcessorTagConfig)
mp.deleteTagsIf = make(map[*vm.Program]messageProcessorTagConfig)
mp.deleteMetaIf = make(map[*vm.Program]messageProcessorTagConfig)
mp.deleteFieldIf = make(map[*vm.Program]messageProcessorTagConfig)
mp.moveFieldToMeta = make(map[*vm.Program]messageProcessorTagConfig)
mp.moveFieldToTag = make(map[*vm.Program]messageProcessorTagConfig)
mp.moveMetaToField = make(map[*vm.Program]messageProcessorTagConfig)
mp.moveMetaToTag = make(map[*vm.Program]messageProcessorTagConfig)
mp.moveTagToField = make(map[*vm.Program]messageProcessorTagConfig)
mp.moveTagToMeta = make(map[*vm.Program]messageProcessorTagConfig)
mp.normalizeUnits = false
return nil
}
func (mp *messageProcessor) AddDropMessagesByName(name string) error {
mp.mutex.Lock()
if _, ok := mp.dropMessages[name]; !ok {
mp.dropMessages[name] = struct{}{}
}
mp.mutex.Unlock()
return nil
}
func (mp *messageProcessor) RemoveDropMessagesByName(name string) {
mp.mutex.Lock()
delete(mp.dropMessages, name)
mp.mutex.Unlock()
}
func (mp *messageProcessor) AddDropMessagesByType(typestring string) error {
valid := []string{"metric", "event", "control", "log"}
isValid := false
for _, t := range valid {
if t == typestring {
isValid = true
break
}
}
if isValid {
mp.mutex.Lock()
if _, ok := mp.dropTypes[typestring]; !ok {
cclog.ComponentDebug("MessageProcessor", "Adding type", typestring, "for dropping")
mp.dropTypes[typestring] = struct{}{}
}
mp.mutex.Unlock()
} else {
return fmt.Errorf("invalid message type %s", typestring)
}
return nil
}
func (mp *messageProcessor) RemoveDropMessagesByType(typestring string) {
mp.mutex.Lock()
delete(mp.dropTypes, typestring)
mp.mutex.Unlock()
}
func (mp *messageProcessor) addTagConfig(condition, key, value string, config *map[*vm.Program]messageProcessorTagConfig) error {
var err error
evaluable, err := expr.Compile(sanitizeExprString(condition), expr.Env(baseenv), expr.AsBool())
if err != nil {
return fmt.Errorf("failed to create condition evaluable of '%s': %v", condition, err.Error())
}
mp.mutex.Lock()
if _, ok := (*config)[evaluable]; !ok {
mp.mapping[condition] = evaluable
(*config)[evaluable] = messageProcessorTagConfig{
Condition: condition,
Key: key,
Value: value,
}
}
mp.mutex.Unlock()
return nil
}
func (mp *messageProcessor) removeTagConfig(condition string, config *map[*vm.Program]messageProcessorTagConfig) {
mp.mutex.Lock()
if e, ok := mp.mapping[condition]; ok {
delete(mp.mapping, condition)
delete(*config, e)
}
mp.mutex.Unlock()
}
func (mp *messageProcessor) AddAddTagsByCondition(condition, key, value string) error {
return mp.addTagConfig(condition, key, value, &mp.addTagsIf)
}
func (mp *messageProcessor) RemoveAddTagsByCondition(condition string) {
mp.removeTagConfig(condition, &mp.addTagsIf)
}
func (mp *messageProcessor) AddDeleteTagsByCondition(condition, key, value string) error {
return mp.addTagConfig(condition, key, value, &mp.deleteTagsIf)
}
func (mp *messageProcessor) RemoveDeleteTagsByCondition(condition string) {
mp.removeTagConfig(condition, &mp.deleteTagsIf)
}
func (mp *messageProcessor) AddAddMetaByCondition(condition, key, value string) error {
return mp.addTagConfig(condition, key, value, &mp.addMetaIf)
}
func (mp *messageProcessor) RemoveAddMetaByCondition(condition string) {
mp.removeTagConfig(condition, &mp.addMetaIf)
}
func (mp *messageProcessor) AddDeleteMetaByCondition(condition, key, value string) error {
return mp.addTagConfig(condition, key, value, &mp.deleteMetaIf)
}
func (mp *messageProcessor) RemoveDeleteMetaByCondition(condition string) {
mp.removeTagConfig(condition, &mp.deleteMetaIf)
}
func (mp *messageProcessor) AddAddFieldByCondition(condition, key, value string) error {
return mp.addTagConfig(condition, key, value, &mp.addFieldIf)
}
func (mp *messageProcessor) RemoveAddFieldByCondition(condition string) {
mp.removeTagConfig(condition, &mp.addFieldIf)
}
func (mp *messageProcessor) AddDeleteFieldByCondition(condition, key, value string) error {
return mp.addTagConfig(condition, key, value, &mp.deleteFieldIf)
}
func (mp *messageProcessor) RemoveDeleteFieldByCondition(condition string) {
mp.removeTagConfig(condition, &mp.deleteFieldIf)
}
func (mp *messageProcessor) AddDropMessagesByCondition(condition string) error {
var err error
evaluable, err := expr.Compile(sanitizeExprString(condition), expr.Env(baseenv), expr.AsBool())
if err != nil {
return fmt.Errorf("failed to create condition evaluable of '%s': %v", condition, err.Error())
}
mp.mutex.Lock()
if _, ok := mp.dropMessagesIf[evaluable]; !ok {
mp.mapping[condition] = evaluable
mp.dropMessagesIf[evaluable] = struct{}{}
}
mp.mutex.Unlock()
return nil
}
func (mp *messageProcessor) RemoveDropMessagesByCondition(condition string) {
mp.mutex.Lock()
if e, ok := mp.mapping[condition]; ok {
delete(mp.mapping, condition)
delete(mp.dropMessagesIf, e)
}
mp.mutex.Unlock()
}
func (mp *messageProcessor) AddRenameMetricByCondition(condition string, name string) error {
var err error
evaluable, err := expr.Compile(sanitizeExprString(condition), expr.Env(baseenv), expr.AsBool())
if err != nil {
return fmt.Errorf("failed to create condition evaluable of '%s': %v", condition, err.Error())
}
mp.mutex.Lock()
if _, ok := mp.renameMessagesIf[evaluable]; !ok {
mp.mapping[condition] = evaluable
mp.renameMessagesIf[evaluable] = name
} else {
mp.renameMessagesIf[evaluable] = name
}
mp.mutex.Unlock()
return nil
}
func (mp *messageProcessor) RemoveRenameMetricByCondition(condition string) {
mp.mutex.Lock()
if e, ok := mp.mapping[condition]; ok {
delete(mp.mapping, condition)
delete(mp.renameMessagesIf, e)
}
mp.mutex.Unlock()
}
func (mp *messageProcessor) SetNormalizeUnits(setting bool) {
mp.normalizeUnits = setting
}
func (mp *messageProcessor) AddChangeUnitPrefix(condition string, prefix string) error {
var err error
evaluable, err := expr.Compile(sanitizeExprString(condition), expr.Env(baseenv), expr.AsBool())
if err != nil {
return fmt.Errorf("failed to create condition evaluable of '%s': %v", condition, err.Error())
}
mp.mutex.Lock()
if _, ok := mp.changeUnitPrefix[evaluable]; !ok {
mp.mapping[condition] = evaluable
mp.changeUnitPrefix[evaluable] = prefix
} else {
mp.changeUnitPrefix[evaluable] = prefix
}
mp.mutex.Unlock()
return nil
}
func (mp *messageProcessor) RemoveChangeUnitPrefix(condition string) {
mp.mutex.Lock()
if e, ok := mp.mapping[condition]; ok {
delete(mp.mapping, condition)
delete(mp.changeUnitPrefix, e)
}
mp.mutex.Unlock()
}
func (mp *messageProcessor) AddRenameMetricByName(from, to string) error {
mp.mutex.Lock()
if _, ok := mp.renameMessages[from]; !ok {
mp.renameMessages[from] = to
}
mp.mutex.Unlock()
return nil
}
func (mp *messageProcessor) RemoveRenameMetricByName(from string) {
mp.mutex.Lock()
delete(mp.renameMessages, from)
mp.mutex.Unlock()
}
func (mp *messageProcessor) AddMoveTagToMeta(condition, key, value string) error {
return mp.addTagConfig(condition, key, value, &mp.moveTagToMeta)
}
func (mp *messageProcessor) RemoveMoveTagToMeta(condition string) {
mp.removeTagConfig(condition, &mp.moveTagToMeta)
}
func (mp *messageProcessor) AddMoveTagToFields(condition, key, value string) error {
return mp.addTagConfig(condition, key, value, &mp.moveTagToField)
}
func (mp *messageProcessor) RemoveMoveTagToFields(condition string) {
mp.removeTagConfig(condition, &mp.moveTagToField)
}
func (mp *messageProcessor) AddMoveMetaToTags(condition, key, value string) error {
return mp.addTagConfig(condition, key, value, &mp.moveMetaToTag)
}
func (mp *messageProcessor) RemoveMoveMetaToTags(condition string) {
mp.removeTagConfig(condition, &mp.moveMetaToTag)
}
func (mp *messageProcessor) AddMoveMetaToFields(condition, key, value string) error {
return mp.addTagConfig(condition, key, value, &mp.moveMetaToField)
}
func (mp *messageProcessor) RemoveMoveMetaToFields(condition string) {
mp.removeTagConfig(condition, &mp.moveMetaToField)
}
func (mp *messageProcessor) AddMoveFieldToTags(condition, key, value string) error {
return mp.addTagConfig(condition, key, value, &mp.moveFieldToTag)
}
func (mp *messageProcessor) RemoveMoveFieldToTags(condition string) {
mp.removeTagConfig(condition, &mp.moveFieldToTag)
}
func (mp *messageProcessor) AddMoveFieldToMeta(condition, key, value string) error {
return mp.addTagConfig(condition, key, value, &mp.moveFieldToMeta)
}
func (mp *messageProcessor) RemoveMoveFieldToMeta(condition string) {
mp.removeTagConfig(condition, &mp.moveFieldToMeta)
}
func (mp *messageProcessor) SetStages(stages []string) error {
newstages := make([]string, 0)
if len(stages) == 0 {
mp.mutex.Lock()
mp.stages = newstages
mp.mutex.Unlock()
return nil
}
for i, s := range stages {
valid := false
for _, v := range StageNames {
if s == v {
valid = true
}
}
if valid {
newstages = append(newstages, s)
} else {
return fmt.Errorf("invalid stage %s at index %d", s, i)
}
}
mp.mutex.Lock()
mp.stages = newstages
mp.mutex.Unlock()
return nil
}
func (mp *messageProcessor) DefaultStages() []string {
return StageNames
}
func (mp *messageProcessor) FromConfigJSON(config json.RawMessage) error {
var c messageProcessorConfig
err := json.Unmarshal(config, &c)
if err != nil {
return fmt.Errorf("failed to process config JSON: %v", err.Error())
}
if len(c.StageOrder) > 0 {
err = mp.SetStages(c.StageOrder)
if err != nil {
return fmt.Errorf("failed to process config JSON: %v", err.Error())
}
} else {
err = mp.SetStages(mp.DefaultStages())
if err != nil {
return fmt.Errorf("failed to process config JSON: %v", err.Error())
}
}
for _, m := range c.DropMessages {
err = mp.AddDropMessagesByName(m)
if err != nil {
return fmt.Errorf("failed to process config JSON: %v", err.Error())
}
}
for _, m := range c.DropByType {
err = mp.AddDropMessagesByType(m)
if err != nil {
return fmt.Errorf("failed to process config JSON: %v", err.Error())
}
}
for _, m := range c.DropMessagesIf {
err = mp.AddDropMessagesByCondition(m)
if err != nil {
return fmt.Errorf("failed to process config JSON: %v", err.Error())
}
}
for k, v := range c.RenameMessagesIf {
err = mp.AddRenameMetricByCondition(k, v)
if err != nil {
return fmt.Errorf("failed to process config JSON: %v", err.Error())
}
}
for k, v := range c.RenameMessages {
err = mp.AddRenameMetricByName(k, v)
if err != nil {
return fmt.Errorf("failed to process config JSON: %v", err.Error())
}
}
for k, v := range c.ChangeUnitPrefix {
err = mp.AddChangeUnitPrefix(k, v)
if err != nil {
return fmt.Errorf("failed to process config JSON: %v", err.Error())
}
}
for _, c := range c.AddTagsIf {
err = mp.AddAddTagsByCondition(c.Condition, c.Key, c.Value)
if err != nil {
return fmt.Errorf("failed to process config JSON: %v", err.Error())
}
}
for _, c := range c.AddMetaIf {
err = mp.AddAddMetaByCondition(c.Condition, c.Key, c.Value)
if err != nil {
return fmt.Errorf("failed to process config JSON: %v", err.Error())
}
}
for _, c := range c.AddFieldIf {
err = mp.AddAddFieldByCondition(c.Condition, c.Key, c.Value)
if err != nil {
return fmt.Errorf("failed to process config JSON: %v", err.Error())
}
}
for _, c := range c.DelTagsIf {
err = mp.AddDeleteTagsByCondition(c.Condition, c.Key, c.Value)
if err != nil {
return fmt.Errorf("failed to process config JSON: %v", err.Error())
}
}
for _, c := range c.DelMetaIf {
err = mp.AddDeleteMetaByCondition(c.Condition, c.Key, c.Value)
if err != nil {
return fmt.Errorf("failed to process config JSON: %v", err.Error())
}
}
for _, c := range c.DelFieldIf {
err = mp.AddDeleteFieldByCondition(c.Condition, c.Key, c.Value)
if err != nil {
return fmt.Errorf("failed to process config JSON: %v", err.Error())
}
}
for _, c := range c.MoveTagToMeta {
err = mp.AddMoveTagToMeta(c.Condition, c.Key, c.Value)
if err != nil {
return fmt.Errorf("failed to process config JSON: %v", err.Error())
}
}
for _, c := range c.MoveTagToField {
err = mp.AddMoveTagToFields(c.Condition, c.Key, c.Value)
if err != nil {
return fmt.Errorf("failed to process config JSON: %v", err.Error())
}
}
for _, c := range c.MoveMetaToTag {
err = mp.AddMoveMetaToTags(c.Condition, c.Key, c.Value)
if err != nil {
return fmt.Errorf("failed to process config JSON: %v", err.Error())
}
}
for _, c := range c.MoveMetaToField {
err = mp.AddMoveMetaToFields(c.Condition, c.Key, c.Value)
if err != nil {
return fmt.Errorf("failed to process config JSON: %v", err.Error())
}
}
for _, c := range c.MoveFieldToTag {
err = mp.AddMoveFieldToTags(c.Condition, c.Key, c.Value)
if err != nil {
return fmt.Errorf("failed to process config JSON: %v", err.Error())
}
}
for _, c := range c.MoveFieldToMeta {
err = mp.AddMoveFieldToMeta(c.Condition, c.Key, c.Value)
if err != nil {
return fmt.Errorf("failed to process config JSON: %v", err.Error())
}
}
for _, m := range c.DropByType {
err = mp.AddDropMessagesByType(m)
if err != nil {
return fmt.Errorf("failed to process config JSON: %v", err.Error())
}
}
if len(c.AddBaseEnv) > 0 {
err = mp.AddBaseEnv(c.AddBaseEnv)
if err != nil {
return fmt.Errorf("failed to process config JSON: %v", err.Error())
}
}
mp.SetNormalizeUnits(c.NormalizeUnits)
return nil
}
func (mp *messageProcessor) ProcessMetric(metric lplegacy.CCMetric) (lp.CCMessage, error) {
m, err := lp.NewMessage(
metric.Name(),
metric.Tags(),
metric.Meta(),
metric.Fields(),
metric.Time(),
)
if err != nil {
return m, fmt.Errorf("failed to parse metric to message: %v", err.Error())
}
return mp.ProcessMessage(m)
}
func (mp *messageProcessor) ProcessMessage(m lp.CCMessage) (lp.CCMessage, error) {
var err error = nil
var out lp.CCMessage = lp.FromMessage(m)
name := out.Name()
if len(mp.stages) == 0 {
mp.SetStages(mp.DefaultStages())
}
mp.mutex.RLock()
defer mp.mutex.RUnlock()
params := getParamMap(out)
defer func() {
params["field"] = nil
params["tag"] = nil
paramMapPool.Put(params["fields"])
paramMapPool.Put(params["tags"])
paramMapPool.Put(params["meta"])
paramMapPool.Put(params)
}()
for _, s := range mp.stages {
switch s {
case STAGENAME_DROP_BY_NAME:
if len(mp.dropMessages) > 0 {
//cclog.ComponentDebug("MessageProcessor", "Dropping by message name ", name)
if _, ok := mp.dropMessages[name]; ok {
//cclog.ComponentDebug("MessageProcessor", "Drop")
return nil, nil
}
}
case STAGENAME_DROP_BY_TYPE:
if len(mp.dropTypes) > 0 {
//cclog.ComponentDebug("MessageProcessor", "Dropping by message type")
if _, ok := mp.dropTypes[params["messagetype"].(string)]; ok {
//cclog.ComponentDebug("MessageProcessor", "Drop")
return nil, nil
}
}
case STAGENAME_DROP_IF:
if len(mp.dropMessagesIf) > 0 {
//cclog.ComponentDebug("MessageProcessor", "Dropping by condition")
drop, err := dropMessagesIf(&params, &mp.dropMessagesIf)
if err != nil {
return out, fmt.Errorf("failed to evaluate: %v", err.Error())
}
if drop {
//cclog.ComponentDebug("MessageProcessor", "Drop")
return nil, nil
}
}
case STAGENAME_RENAME_BY_NAME:
if len(mp.renameMessages) > 0 {
//cclog.ComponentDebug("MessageProcessor", "Renaming by name match")
if newname, ok := mp.renameMessages[name]; ok {
//cclog.ComponentDebug("MessageProcessor", "Rename to", newname)
out.SetName(newname)
//cclog.ComponentDebug("MessageProcessor", "Add old name as 'oldname' to meta", name)
out.AddMeta("oldname", name)
}
}
case STAGENAME_RENAME_IF:
if len(mp.renameMessagesIf) > 0 {
//cclog.ComponentDebug("MessageProcessor", "Renaming by condition")
_, err := renameMessagesIf(out, &params, &mp.renameMessagesIf)
if err != nil {
return out, fmt.Errorf("failed to evaluate: %v", err.Error())
}
}
case STAGENAME_ADD_TAG:
if len(mp.addTagsIf) > 0 {
//cclog.ComponentDebug("MessageProcessor", "Adding tags")
_, err = addTagIf(out, &params, &mp.addTagsIf)
if err != nil {
return out, fmt.Errorf("failed to evaluate: %v", err.Error())
}
}
case STAGENAME_DELETE_TAG:
if len(mp.deleteTagsIf) > 0 {
//cclog.ComponentDebug("MessageProcessor", "Delete tags")
_, err = deleteTagIf(out, &params, &mp.deleteTagsIf)
if err != nil {
return out, fmt.Errorf("failed to evaluate: %v", err.Error())
}
}
case STAGENAME_ADD_META:
if len(mp.addMetaIf) > 0 {
//cclog.ComponentDebug("MessageProcessor", "Adding meta information")
_, err = addMetaIf(out, &params, &mp.addMetaIf)
if err != nil {
return out, fmt.Errorf("failed to evaluate: %v", err.Error())
}
}
case STAGENAME_DELETE_META:
if len(mp.deleteMetaIf) > 0 {
//cclog.ComponentDebug("MessageProcessor", "Delete meta information")
_, err = deleteMetaIf(out, &params, &mp.deleteMetaIf)
if err != nil {
return out, fmt.Errorf("failed to evaluate: %v", err.Error())
}
}
case STAGENAME_ADD_FIELD:
if len(mp.addFieldIf) > 0 {
//cclog.ComponentDebug("MessageProcessor", "Adding fields")
_, err = addFieldIf(out, &params, &mp.addFieldIf)
if err != nil {
return out, fmt.Errorf("failed to evaluate: %v", err.Error())
}
}
case STAGENAME_DELETE_FIELD:
if len(mp.deleteFieldIf) > 0 {
//cclog.ComponentDebug("MessageProcessor", "Delete fields")
_, err = deleteFieldIf(out, &params, &mp.deleteFieldIf)
if err != nil {
return out, fmt.Errorf("failed to evaluate: %v", err.Error())
}
}
case STAGENAME_MOVE_TAG_META:
if len(mp.moveTagToMeta) > 0 {
//cclog.ComponentDebug("MessageProcessor", "Move tag to meta")
_, err := moveTagToMeta(out, &params, &mp.moveTagToMeta)
if err != nil {
return out, fmt.Errorf("failed to evaluate: %v", err.Error())
}
}
case STAGENAME_MOVE_TAG_FIELD:
if len(mp.moveTagToField) > 0 {
//cclog.ComponentDebug("MessageProcessor", "Move tag to fields")
_, err := moveTagToField(out, &params, &mp.moveTagToField)
if err != nil {
return out, fmt.Errorf("failed to evaluate: %v", err.Error())
}
}
case STAGENAME_MOVE_META_TAG:
if len(mp.moveMetaToTag) > 0 {
//cclog.ComponentDebug("MessageProcessor", "Move meta to tags")
_, err := moveMetaToTag(out, &params, &mp.moveMetaToTag)
if err != nil {
return out, fmt.Errorf("failed to evaluate: %v", err.Error())
}
}
case STAGENAME_MOVE_META_FIELD:
if len(mp.moveMetaToField) > 0 {
//cclog.ComponentDebug("MessageProcessor", "Move meta to fields")
_, err := moveMetaToField(out, &params, &mp.moveMetaToField)
if err != nil {
return out, fmt.Errorf("failed to evaluate: %v", err.Error())
}
}
case STAGENAME_MOVE_FIELD_META:
if len(mp.moveFieldToMeta) > 0 {
//cclog.ComponentDebug("MessageProcessor", "Move field to meta")
_, err := moveFieldToMeta(out, &params, &mp.moveFieldToMeta)
if err != nil {
return out, fmt.Errorf("failed to evaluate: %v", err.Error())
}
}
case STAGENAME_MOVE_FIELD_TAG:
if len(mp.moveFieldToTag) > 0 {
//cclog.ComponentDebug("MessageProcessor", "Move field to tags")
_, err := moveFieldToTag(out, &params, &mp.moveFieldToTag)
if err != nil {
return out, fmt.Errorf("failed to evaluate: %v", err.Error())
}
}
case STAGENAME_NORMALIZE_UNIT:
if mp.normalizeUnits {
//cclog.ComponentDebug("MessageProcessor", "Normalize units")
if lp.IsMetric(out) {
_, err := normalizeUnits(out)
if err != nil {
return out, fmt.Errorf("failed to evaluate: %v", err.Error())
}
} else {
cclog.ComponentDebug("MessageProcessor", "skipped, no metric")
}
}
case STAGENAME_CHANGE_UNIT_PREFIX:
if len(mp.changeUnitPrefix) > 0 {
//cclog.ComponentDebug("MessageProcessor", "Change unit prefix")
if lp.IsMetric(out) {
_, err := changeUnitPrefix(out, &params, &mp.changeUnitPrefix)
if err != nil {
return out, fmt.Errorf("failed to evaluate: %v", err.Error())
}
} else {
cclog.ComponentDebug("MessageProcessor", "skipped, no metric")
}
}
}
}
return out, nil
}
// Get a new instace of a message processor.
func NewMessageProcessor() (MessageProcessor, error) {
mp := new(messageProcessor)
err := mp.init()
if err != nil {
err := fmt.Errorf("failed to create MessageProcessor: %v", err.Error())
cclog.ComponentError("MessageProcessor", err.Error())
return nil, err
}
return mp, nil
}

View File

@@ -0,0 +1,262 @@
package messageprocessor
import (
"errors"
"fmt"
lp2 "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
units "github.com/ClusterCockpit/cc-units"
"github.com/expr-lang/expr"
"github.com/expr-lang/expr/vm"
)
type MessageLocation int
const (
MESSAGE_LOCATION_TAGS MessageLocation = iota
MESSAGE_LOCATION_META
MESSAGE_LOCATION_FIELDS
)
// Abstract function to move entries from one location to another
func moveInMessage(message lp2.CCMessage, params *map[string]interface{}, checks *map[*vm.Program]messageProcessorTagConfig, from, to MessageLocation) (bool, error) {
for d, data := range *checks {
value, err := expr.Run(d, *params)
if err != nil {
return false, fmt.Errorf("failed to evaluate: %v", err.Error())
}
//cclog.ComponentDebug("MessageProcessor", "Move from", from, "to", to)
if value.(bool) {
var v string
var ok bool = false
switch from {
case MESSAGE_LOCATION_TAGS:
//cclog.ComponentDebug("MessageProcessor", "Getting tag key", data.Key)
v, ok = message.GetTag(data.Key)
case MESSAGE_LOCATION_META:
//cclog.ComponentDebug("MessageProcessor", "Getting meta key", data.Key)
//cclog.ComponentDebug("MessageProcessor", message.Meta())
v, ok = message.GetMeta(data.Key)
case MESSAGE_LOCATION_FIELDS:
var x interface{}
//cclog.ComponentDebug("MessageProcessor", "Getting field key", data.Key)
x, ok = message.GetField(data.Key)
v = fmt.Sprintf("%v", x)
}
if ok {
switch from {
case MESSAGE_LOCATION_TAGS:
//cclog.ComponentDebug("MessageProcessor", "Removing tag key", data.Key)
message.RemoveTag(data.Key)
case MESSAGE_LOCATION_META:
//cclog.ComponentDebug("MessageProcessor", "Removing meta key", data.Key)
message.RemoveMeta(data.Key)
case MESSAGE_LOCATION_FIELDS:
//cclog.ComponentDebug("MessageProcessor", "Removing field key", data.Key)
message.RemoveField(data.Key)
}
switch to {
case MESSAGE_LOCATION_TAGS:
//cclog.ComponentDebug("MessageProcessor", "Adding tag", data.Value, "->", v)
message.AddTag(data.Value, v)
case MESSAGE_LOCATION_META:
//cclog.ComponentDebug("MessageProcessor", "Adding meta", data.Value, "->", v)
message.AddMeta(data.Value, v)
case MESSAGE_LOCATION_FIELDS:
//cclog.ComponentDebug("MessageProcessor", "Adding field", data.Value, "->", v)
message.AddField(data.Value, v)
}
}
}
}
return false, nil
}
func deleteIf(message lp2.CCMessage, params *map[string]interface{}, checks *map[*vm.Program]messageProcessorTagConfig, location MessageLocation) (bool, error) {
for d, data := range *checks {
value, err := expr.Run(d, *params)
if err != nil {
return true, fmt.Errorf("failed to evaluate: %v", err.Error())
}
if value.(bool) {
switch location {
case MESSAGE_LOCATION_FIELDS:
switch data.Key {
case "value", "event", "log", "control":
return false, errors.New("cannot delete protected fields")
default:
//cclog.ComponentDebug("MessageProcessor", "Removing field for", data.Key)
message.RemoveField(data.Key)
}
case MESSAGE_LOCATION_TAGS:
//cclog.ComponentDebug("MessageProcessor", "Removing tag for", data.Key)
message.RemoveTag(data.Key)
case MESSAGE_LOCATION_META:
//cclog.ComponentDebug("MessageProcessor", "Removing meta for", data.Key)
message.RemoveMeta(data.Key)
}
}
}
return false, nil
}
func addIf(message lp2.CCMessage, params *map[string]interface{}, checks *map[*vm.Program]messageProcessorTagConfig, location MessageLocation) (bool, error) {
for d, data := range *checks {
value, err := expr.Run(d, *params)
if err != nil {
return true, fmt.Errorf("failed to evaluate: %v", err.Error())
}
if value.(bool) {
switch location {
case MESSAGE_LOCATION_FIELDS:
//cclog.ComponentDebug("MessageProcessor", "Adding field", data.Value, "->", data.Value)
message.AddField(data.Key, data.Value)
case MESSAGE_LOCATION_TAGS:
//cclog.ComponentDebug("MessageProcessor", "Adding tag", data.Value, "->", data.Value)
message.AddTag(data.Key, data.Value)
case MESSAGE_LOCATION_META:
//cclog.ComponentDebug("MessageProcessor", "Adding meta", data.Value, "->", data.Value)
message.AddMeta(data.Key, data.Value)
}
}
}
return false, nil
}
func deleteTagIf(message lp2.CCMessage, params *map[string]interface{}, checks *map[*vm.Program]messageProcessorTagConfig) (bool, error) {
return deleteIf(message, params, checks, MESSAGE_LOCATION_TAGS)
}
func addTagIf(message lp2.CCMessage, params *map[string]interface{}, checks *map[*vm.Program]messageProcessorTagConfig) (bool, error) {
return addIf(message, params, checks, MESSAGE_LOCATION_TAGS)
}
func moveTagToMeta(message lp2.CCMessage, params *map[string]interface{}, checks *map[*vm.Program]messageProcessorTagConfig) (bool, error) {
return moveInMessage(message, params, checks, MESSAGE_LOCATION_TAGS, MESSAGE_LOCATION_META)
}
func moveTagToField(message lp2.CCMessage, params *map[string]interface{}, checks *map[*vm.Program]messageProcessorTagConfig) (bool, error) {
return moveInMessage(message, params, checks, MESSAGE_LOCATION_TAGS, MESSAGE_LOCATION_FIELDS)
}
func deleteMetaIf(message lp2.CCMessage, params *map[string]interface{}, checks *map[*vm.Program]messageProcessorTagConfig) (bool, error) {
return deleteIf(message, params, checks, MESSAGE_LOCATION_META)
}
func addMetaIf(message lp2.CCMessage, params *map[string]interface{}, checks *map[*vm.Program]messageProcessorTagConfig) (bool, error) {
return addIf(message, params, checks, MESSAGE_LOCATION_META)
}
func moveMetaToTag(message lp2.CCMessage, params *map[string]interface{}, checks *map[*vm.Program]messageProcessorTagConfig) (bool, error) {
return moveInMessage(message, params, checks, MESSAGE_LOCATION_META, MESSAGE_LOCATION_TAGS)
}
func moveMetaToField(message lp2.CCMessage, params *map[string]interface{}, checks *map[*vm.Program]messageProcessorTagConfig) (bool, error) {
return moveInMessage(message, params, checks, MESSAGE_LOCATION_META, MESSAGE_LOCATION_FIELDS)
}
func deleteFieldIf(message lp2.CCMessage, params *map[string]interface{}, checks *map[*vm.Program]messageProcessorTagConfig) (bool, error) {
return deleteIf(message, params, checks, MESSAGE_LOCATION_FIELDS)
}
func addFieldIf(message lp2.CCMessage, params *map[string]interface{}, checks *map[*vm.Program]messageProcessorTagConfig) (bool, error) {
return addIf(message, params, checks, MESSAGE_LOCATION_FIELDS)
}
func moveFieldToTag(message lp2.CCMessage, params *map[string]interface{}, checks *map[*vm.Program]messageProcessorTagConfig) (bool, error) {
return moveInMessage(message, params, checks, MESSAGE_LOCATION_FIELDS, MESSAGE_LOCATION_TAGS)
}
func moveFieldToMeta(message lp2.CCMessage, params *map[string]interface{}, checks *map[*vm.Program]messageProcessorTagConfig) (bool, error) {
return moveInMessage(message, params, checks, MESSAGE_LOCATION_FIELDS, MESSAGE_LOCATION_META)
}
func dropMessagesIf(params *map[string]interface{}, checks *map[*vm.Program]struct{}) (bool, error) {
for d := range *checks {
value, err := expr.Run(d, *params)
if err != nil {
return false, fmt.Errorf("failed to evaluate: %v", err.Error())
}
if value.(bool) {
return true, nil
}
}
return false, nil
}
func normalizeUnits(message lp2.CCMessage) (bool, error) {
if in_unit, ok := message.GetMeta("unit"); ok {
u := units.NewUnit(in_unit)
if u.Valid() {
//cclog.ComponentDebug("MessageProcessor", "Update unit with", u.Short())
message.AddMeta("unit", u.Short())
}
} else if in_unit, ok := message.GetTag("unit"); ok {
u := units.NewUnit(in_unit)
if u.Valid() {
//cclog.ComponentDebug("MessageProcessor", "Update unit with", u.Short())
message.AddTag("unit", u.Short())
}
}
return false, nil
}
func changeUnitPrefix(message lp2.CCMessage, params *map[string]interface{}, checks *map[*vm.Program]string) (bool, error) {
for r, n := range *checks {
value, err := expr.Run(r, *params)
if err != nil {
return false, fmt.Errorf("failed to evaluate: %v", err.Error())
}
if value.(bool) {
newPrefix := units.NewPrefix(n)
//cclog.ComponentDebug("MessageProcessor", "Condition matches, change to prefix", newPrefix.String())
if in_unit, ok := message.GetMeta("unit"); ok && newPrefix != units.InvalidPrefix {
u := units.NewUnit(in_unit)
if u.Valid() {
//cclog.ComponentDebug("MessageProcessor", "Input unit", u.Short())
conv, out_unit := units.GetUnitPrefixFactor(u, newPrefix)
if conv != nil && out_unit.Valid() {
if val, ok := message.GetField("value"); ok {
//cclog.ComponentDebug("MessageProcessor", "Update unit with", out_unit.Short())
message.AddField("value", conv(val))
message.AddMeta("unit", out_unit.Short())
}
}
}
} else if in_unit, ok := message.GetTag("unit"); ok && newPrefix != units.InvalidPrefix {
u := units.NewUnit(in_unit)
if u.Valid() {
//cclog.ComponentDebug("MessageProcessor", "Input unit", u.Short())
conv, out_unit := units.GetUnitPrefixFactor(u, newPrefix)
if conv != nil && out_unit.Valid() {
if val, ok := message.GetField("value"); ok {
//cclog.ComponentDebug("MessageProcessor", "Update unit with", out_unit.Short())
message.AddField("value", conv(val))
message.AddTag("unit", out_unit.Short())
}
}
}
}
}
}
return false, nil
}
func renameMessagesIf(message lp2.CCMessage, params *map[string]interface{}, checks *map[*vm.Program]string) (bool, error) {
for d, n := range *checks {
value, err := expr.Run(d, *params)
if err != nil {
return true, fmt.Errorf("failed to evaluate: %v", err.Error())
}
if value.(bool) {
old := message.Name()
//cclog.ComponentDebug("MessageProcessor", "Rename to", n)
message.SetName(n)
//cclog.ComponentDebug("MessageProcessor", "Add old name as 'oldname' to meta", old)
message.AddMeta("oldname", old)
}
}
return false, nil
}

View File

@@ -0,0 +1,396 @@
package messageprocessor
import (
"encoding/json"
"errors"
"fmt"
"testing"
"time"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
cclog "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
)
func generate_message_lists(num_lists, num_entries int) ([][]lp.CCMessage, error) {
mlist := make([][]lp.CCMessage, 0)
for j := 0; j < num_lists; j++ {
out := make([]lp.CCMessage, 0)
for i := 0; i < num_entries; i++ {
var x lp.CCMessage
var err error = nil
switch {
case i%4 == 0:
x, err = lp.NewEvent("myevent", map[string]string{"type": "socket", "type-id": "0"}, map[string]string{}, "nothing happend", time.Now())
case i%4 == 1:
x, err = lp.NewMetric("mymetric", map[string]string{"type": "socket", "type-id": "0"}, map[string]string{"unit": "kByte"}, 12.145, time.Now())
case i%4 == 2:
x, err = lp.NewLog("mylog", map[string]string{"type": "socket", "type-id": "0"}, map[string]string{}, "disk status: OK", time.Now())
case i%4 == 3:
x, err = lp.NewGetControl("mycontrol", map[string]string{"type": "socket", "type-id": "0"}, map[string]string{}, time.Now())
}
if err == nil {
x.AddTag("hostname", "myhost")
out = append(out, x)
} else {
return nil, errors.New("failed to create message")
}
}
mlist = append(mlist, out)
}
return mlist, nil
}
func TestNewMessageProcessor(t *testing.T) {
_, err := NewMessageProcessor()
if err != nil {
t.Error(err.Error())
}
}
type Configs struct {
name string
config json.RawMessage
drop bool
errors bool
pre func(msg lp.CCMessage) error
check func(msg lp.CCMessage) error
}
var test_configs = []Configs{
{
name: "single_dropif_nomatch",
config: json.RawMessage(`{"drop_messages_if": [ "name == 'testname' && tags.type == 'socket' && tags.typeid % 2 == 1"]}`),
},
{
name: "drop_by_name",
config: json.RawMessage(`{"drop_messages": [ "net_bytes_in"]}`),
drop: true,
},
{
name: "drop_by_type_match",
config: json.RawMessage(`{"drop_by_message_type": [ "metric"]}`),
drop: true,
},
{
name: "drop_by_type_nomatch",
config: json.RawMessage(`{"drop_by_message_type": [ "event"]}`),
},
{
name: "single_dropif_match",
config: json.RawMessage(`{"drop_messages_if": [ "name == 'net_bytes_in' && tags.type == 'node'"]}`),
drop: true,
},
{
name: "double_dropif_match_nomatch",
config: json.RawMessage(`{"drop_messages_if": [ "name == 'net_bytes_in' && tags.type == 'node'", "name == 'testname' && tags.type == 'socket' && tags.typeid % 2 == 1"]}`),
drop: true,
},
{
name: "rename_simple",
config: json.RawMessage(`{"rename_messages": { "net_bytes_in" : "net_bytes_out", "rapl_power": "cpu_power"}}`),
check: func(msg lp.CCMessage) error {
if msg.Name() != "net_bytes_out" {
return errors.New("expected name net_bytes_out but still have net_bytes_in")
}
return nil
},
},
{
name: "rename_match",
config: json.RawMessage(`{"rename_messages_if": { "name == 'net_bytes_in'" : "net_bytes_out", "name == 'rapl_power'": "cpu_power"}}`),
check: func(msg lp.CCMessage) error {
if msg.Name() != "net_bytes_out" {
return errors.New("expected name net_bytes_out but still have net_bytes_in")
}
return nil
},
},
{
name: "rename_nomatch",
config: json.RawMessage(`{"rename_messages_if": { "name == 'net_bytes_out'" : "net_bytes_in", "name == 'rapl_power'": "cpu_power"}}`),
check: func(msg lp.CCMessage) error {
if msg.Name() != "net_bytes_in" {
return errors.New("expected name net_bytes_in but still have net_bytes_out")
}
return nil
},
},
{
name: "add_tag",
config: json.RawMessage(`{"add_tags_if": [{"if": "name == 'net_bytes_in'", "key" : "cluster", "value" : "mycluster"}]}`),
check: func(msg lp.CCMessage) error {
if !msg.HasTag("cluster") {
return errors.New("expected new tag 'cluster' but not present")
}
return nil
},
},
{
name: "del_tag",
config: json.RawMessage(`{"delete_tags_if": [{"if": "name == 'net_bytes_in'", "key" : "type"}]}`),
check: func(msg lp.CCMessage) error {
if msg.HasTag("type") {
return errors.New("expected to have no 'type' but still present")
}
return nil
},
},
{
name: "add_meta",
config: json.RawMessage(`{"add_meta_if": [{"if": "name == 'net_bytes_in'", "key" : "source", "value" : "example"}]}`),
check: func(msg lp.CCMessage) error {
if !msg.HasMeta("source") {
return errors.New("expected new tag 'source' but not present")
}
return nil
},
},
{
name: "del_meta",
config: json.RawMessage(`{"delete_meta_if": [{"if": "name == 'net_bytes_in'", "key" : "unit"}]}`),
check: func(msg lp.CCMessage) error {
if msg.HasMeta("unit") {
return errors.New("expected to have no 'unit' but still present")
}
return nil
},
},
{
name: "add_field",
config: json.RawMessage(`{"add_fields_if": [{"if": "name == 'net_bytes_in'", "key" : "myfield", "value" : "example"}]}`),
check: func(msg lp.CCMessage) error {
if !msg.HasField("myfield") {
return errors.New("expected new tag 'source' but not present")
}
return nil
},
},
{
name: "delete_fields_if_protected",
config: json.RawMessage(`{"delete_fields_if": [{"if": "name == 'net_bytes_in'", "key" : "value"}]}`),
errors: true,
check: func(msg lp.CCMessage) error {
if !msg.HasField("value") {
return errors.New("expected to still have 'value' field because it is a protected field key")
}
return nil
},
},
{
name: "delete_fields_if_unprotected",
config: json.RawMessage(`{"delete_fields_if": [{"if": "name == 'net_bytes_in'", "key" : "testfield"}]}`),
check: func(msg lp.CCMessage) error {
if msg.HasField("testfield") {
return errors.New("expected to still have 'testfield' field but should be deleted")
}
return nil
},
pre: func(msg lp.CCMessage) error {
msg.AddField("testfield", 4.123)
return nil
},
},
{
name: "single_change_prefix_match",
config: json.RawMessage(`{"change_unit_prefix": {"name == 'net_bytes_in' && tags.type == 'node'": "M"}}`),
check: func(msg lp.CCMessage) error {
if u, ok := msg.GetMeta("unit"); ok {
if u != "MB" {
return fmt.Errorf("expected unit MB but have %s", u)
}
} else if u, ok := msg.GetTag("unit"); ok {
if u != "MB" {
return fmt.Errorf("expected unit MB but have %s", u)
}
}
return nil
},
},
{
name: "normalize_units",
config: json.RawMessage(`{"normalize_units": true}`),
check: func(msg lp.CCMessage) error {
if u, ok := msg.GetMeta("unit"); ok {
if u != "B" {
return fmt.Errorf("expected unit B but have %s", u)
}
} else if u, ok := msg.GetTag("unit"); ok {
if u != "B" {
return fmt.Errorf("expected unit B but have %s", u)
}
}
return nil
},
},
{
name: "move_tag_to_meta",
config: json.RawMessage(`{"move_tag_to_meta_if": [{"if": "name == 'net_bytes_in'", "key" : "type-id", "value": "typeid"}]}`),
check: func(msg lp.CCMessage) error {
if msg.HasTag("type-id") || !msg.HasMeta("typeid") {
return errors.New("moving tag 'type-id' to meta 'typeid' failed")
}
return nil
},
pre: func(msg lp.CCMessage) error {
msg.AddTag("type-id", "0")
return nil
},
},
{
name: "move_tag_to_field",
config: json.RawMessage(`{"move_tag_to_field_if": [{"if": "name == 'net_bytes_in'", "key" : "type-id", "value": "typeid"}]}`),
check: func(msg lp.CCMessage) error {
if msg.HasTag("type-id") || !msg.HasField("typeid") {
return errors.New("moving tag 'type-id' to field 'typeid' failed")
}
return nil
},
pre: func(msg lp.CCMessage) error {
msg.AddTag("type-id", "0")
return nil
},
},
{
name: "move_meta_to_tag",
config: json.RawMessage(`{"move_meta_to_tag_if": [{"if": "name == 'net_bytes_in'", "key" : "unit", "value": "unit"}]}`),
check: func(msg lp.CCMessage) error {
if msg.HasMeta("unit") || !msg.HasTag("unit") {
return errors.New("moving meta 'unit' to tag 'unit' failed")
}
return nil
},
},
{
name: "move_meta_to_field",
config: json.RawMessage(`{"move_meta_to_field_if": [{"if": "name == 'net_bytes_in'", "key" : "unit", "value": "unit"}]}`),
check: func(msg lp.CCMessage) error {
if msg.HasMeta("unit") || !msg.HasField("unit") {
return errors.New("moving meta 'unit' to field 'unit' failed")
}
return nil
},
},
{
name: "move_field_to_tag",
config: json.RawMessage(`{"move_field_to_tag_if": [{"if": "name == 'net_bytes_in'", "key" : "myfield", "value": "field"}]}`),
check: func(msg lp.CCMessage) error {
if msg.HasField("myfield") || !msg.HasTag("field") {
return errors.New("moving meta 'myfield' to tag 'field' failed")
}
return nil
},
pre: func(msg lp.CCMessage) error {
msg.AddField("myfield", 12)
return nil
},
},
{
name: "move_field_to_meta",
config: json.RawMessage(`{"move_field_to_meta_if": [{"if": "name == 'net_bytes_in'", "key" : "myfield", "value": "field"}]}`),
check: func(msg lp.CCMessage) error {
if msg.HasField("myfield") || !msg.HasMeta("field") {
return errors.New("moving meta 'myfield' to meta 'field' failed")
}
return nil
},
pre: func(msg lp.CCMessage) error {
msg.AddField("myfield", 12)
return nil
},
},
}
func TestConfigList(t *testing.T) {
for _, c := range test_configs {
t.Run(c.name, func(t *testing.T) {
m, err := lp.NewMetric("net_bytes_in", map[string]string{"type": "node", "type-id": "0"}, map[string]string{"unit": "Byte"}, float64(1024.0), time.Now())
if err != nil {
t.Error(err.Error())
return
}
if c.pre != nil {
if err = c.pre(m); err != nil {
t.Errorf("error running pre-test function: %v", err.Error())
return
}
}
mp, err := NewMessageProcessor()
if err != nil {
t.Error(err.Error())
return
}
err = mp.FromConfigJSON(c.config)
if err != nil {
t.Error(err.Error())
return
}
//t.Log(m.ToLineProtocol(nil))
out, err := mp.ProcessMessage(m)
if err != nil && !c.errors {
cclog.SetDebug()
mp.ProcessMessage(m)
t.Error(err.Error())
return
}
if out == nil && !c.drop {
t.Error("fail, message should NOT be dropped but processor signalled dropping")
return
} else if out != nil && c.drop {
t.Error("fail, message should be dropped but processor signalled NO dropping")
return
}
// {
// if c.drop {
// t.Error("fail, message should be dropped but processor signalled NO dropping")
// } else {
// t.Error("fail, message should NOT be dropped but processor signalled dropping")
// }
// cclog.SetDebug()
// mp.ProcessMessage(m)
// return
// }
if c.check != nil {
if err := c.check(out); err != nil {
t.Errorf("check failed with %v", err.Error())
t.Log("Rerun with debugging")
cclog.SetDebug()
mp.ProcessMessage(m)
return
}
}
})
}
}
func BenchmarkProcessing(b *testing.B) {
mlist, err := generate_message_lists(b.N, 1000)
if err != nil {
b.Error(err.Error())
return
}
mp, err := NewMessageProcessor()
if err != nil {
b.Error(err.Error())
return
}
err = mp.FromConfigJSON(json.RawMessage(`{"move_meta_to_tag_if": [{"if" : "name == 'mymetric'", "key":"unit", "value":"unit"}]}`))
if err != nil {
b.Error(err.Error())
return
}
b.StartTimer()
for i := 0; i < b.N; i++ {
for _, m := range mlist[i] {
if _, err := mp.ProcessMessage(m); err != nil {
b.Errorf("failed processing message '%s': %v", m.ToLineProtocol(nil), err.Error())
return
}
}
}
b.StopTimer()
b.ReportMetric(float64(b.Elapsed())/float64(len(mlist)*b.N), "ns/message")
}

View File

@@ -1,14 +1,3 @@
<!--
---
title: Multi-channel Ticker
description: Timer ticker that sends out the tick to multiple channels
categories: [cc-metric-collector]
tags: ['Developer']
weight: 1
hugo_path: docs/reference/cc-metric-collector/pkg/multichanticker/_index.md
---
-->
# MultiChanTicker
The idea of this ticker is to multiply the output channels. The original Golang `time.Ticker` provides only a single output channel, so the signal can only be received by a single other class. This ticker allows to add multiple channels which get all notified about the time tick.

View File

@@ -3,7 +3,7 @@ package multiChanTicker
import (
"time"
cclog "github.com/ClusterCockpit/cc-lib/ccLogger"
cclog "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
)
type multiChanTicker struct {

32
receivers/README.md Normal file
View File

@@ -0,0 +1,32 @@
# CCMetric receivers
This folder contains the ReceiveManager and receiver implementations for the cc-metric-collector.
## Configuration
The configuration file for the receivers is a list of configurations. The `type` field in each specifies which receiver to initialize.
```json
{
"myreceivername" : {
"type": "receiver-type",
<receiver-specific configuration>
}
}
```
This allows to specify
## Available receivers
- [`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
- [`ipmi`](./ipmiReceiver.md): Read IPMI sensor readings
- [`redfish`](redfishReceiver.md) Use the Redfish (specification) to query thermal and power metrics
## Contributing own receivers
A receiver contains a few functions and is derived from the type `Receiver` (in `metricReceiver.go`):
For an example, check the [sample receiver](./sampleReceiver.go)

183
receivers/httpReceiver.go Normal file
View File

@@ -0,0 +1,183 @@
package receivers
import (
"context"
"encoding/json"
"errors"
"fmt"
"net/http"
"strings"
"sync"
"time"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
cclog "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
mp "github.com/ClusterCockpit/cc-metric-collector/pkg/messageProcessor"
)
const HTTP_RECEIVER_PORT = "8080"
type HttpReceiverConfig struct {
defaultReceiverConfig
Addr string `json:"address"`
Port string `json:"port"`
Path string `json:"path"`
// Maximum amount of time to wait for the next request when keep-alives are enabled
// should be larger than the measurement interval to keep the connection open
IdleTimeout string `json:"idle_timeout"`
idleTimeout time.Duration
// Controls whether HTTP keep-alives are enabled. By default, keep-alives are enabled
KeepAlivesEnabled bool `json:"keep_alives_enabled"`
// Basic authentication
Username string `json:"username"`
Password string `json:"password"`
useBasicAuth bool
}
type HttpReceiver struct {
receiver
//meta map[string]string
config HttpReceiverConfig
server *http.Server
wg sync.WaitGroup
}
func (r *HttpReceiver) Init(name string, config json.RawMessage) error {
r.name = fmt.Sprintf("HttpReceiver(%s)", name)
// Set default values
r.config.Port = HTTP_RECEIVER_PORT
r.config.KeepAlivesEnabled = true
// should be larger than the measurement interval to keep the connection open
r.config.IdleTimeout = "120s"
// Read config
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")
}
// Check idle timeout config
if len(r.config.IdleTimeout) > 0 {
t, err := time.ParseDuration(r.config.IdleTimeout)
if err == nil {
cclog.ComponentDebug(r.name, "idleTimeout", t)
r.config.idleTimeout = t
}
}
// Check basic authentication config
if len(r.config.Username) > 0 || len(r.config.Password) > 0 {
r.config.useBasicAuth = true
}
if r.config.useBasicAuth && len(r.config.Username) == 0 {
return errors.New("basic authentication requires username")
}
if r.config.useBasicAuth && len(r.config.Password) == 0 {
return errors.New("basic authentication requires password")
}
msgp, err := mp.NewMessageProcessor()
if err != nil {
return fmt.Errorf("initialization of message processor failed: %v", err.Error())
}
r.mp = msgp
if len(r.config.MessageProcessor) > 0 {
err = r.mp.FromConfigJSON(r.config.MessageProcessor)
if err != nil {
return fmt.Errorf("failed parsing JSON for message processor: %v", err.Error())
}
}
r.mp.AddAddMetaByCondition("true", "source", r.name)
//r.meta = map[string]string{"source": r.name}
p := r.config.Path
if !strings.HasPrefix(p, "/") {
p = "/" + p
}
addr := fmt.Sprintf("%s:%s", r.config.Addr, r.config.Port)
uri := addr + p
cclog.ComponentDebug(r.name, "INIT", "listen on:", uri)
// Register handler function r.ServerHttp for path p in the DefaultServeMux
http.HandleFunc(p, r.ServerHttp)
// Create http server
r.server = &http.Server{
Addr: addr,
Handler: nil, // handler to invoke, http.DefaultServeMux if nil
IdleTimeout: r.config.idleTimeout,
}
r.server.SetKeepAlivesEnabled(r.config.KeepAlivesEnabled)
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) {
// Check request method, only post method is handled
if req.Method != http.MethodPost {
http.Error(w, "Method Not Allowed", http.StatusMethodNotAllowed)
return
}
// Check basic authentication
if r.config.useBasicAuth {
username, password, ok := req.BasicAuth()
if !ok || username != r.config.Username || password != r.config.Password {
http.Error(w, "Unauthorized", http.StatusUnauthorized)
return
}
}
if r.sink != nil {
buf := make([]byte, 0, req.ContentLength)
len, err := req.Body.Read(buf)
if err == nil && len > 0 {
messages, err := lp.FromBytes(buf)
if err != nil {
msg := "ServerHttp: Failed to decode messages: " + err.Error()
cclog.ComponentError(r.name, msg)
http.Error(w, msg, http.StatusInternalServerError)
return
}
for _, y := range messages {
m, err := r.mp.ProcessMessage(y)
if err == nil && m != nil {
r.sink <- m
}
}
}
}
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
}

43
receivers/httpReceiver.md Normal file
View File

@@ -0,0 +1,43 @@
## `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",
"idle_timeout": "120s",
"username": "myUser",
"password": "myPW"
}
}
```
- `type`: makes the receiver a `http` receiver
- `address`: Listen address
- `port`: Listen port
- `path`: URL path for the write endpoint
- `idle_timeout`: Maximum amount of time to wait for the next request when keep-alives are enabled should be larger than the measurement interval to keep the connection open
- `keep_alives_enabled`: Controls whether HTTP keep-alives are enabled. By default, keep-alives are enabled.
- `username`: username for basic authentication
- `password`: password for basic authentication
The HTTP endpoint listens to `http://<address>:<port>/<path>`
### Debugging
- Install [curl](https://curl.se/)
- Use curl to send message to `http` receiver
```bash
curl http://localhost:8080/write \
--user "myUser:myPW" \
--data \
"myMetric,hostname=myHost,type=hwthread,type-id=0,unit=Hz value=400000i 1694777161164284635
myMetric,hostname=myHost,type=hwthread,type-id=1,unit=Hz value=400001i 1694777161164284635"
```

571
receivers/ipmiReceiver.go Normal file
View File

@@ -0,0 +1,571 @@
package receivers
import (
"bufio"
"bytes"
"encoding/json"
"fmt"
"io"
"os/exec"
"regexp"
"strconv"
"strings"
"sync"
"time"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
cclog "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
"github.com/ClusterCockpit/cc-metric-collector/pkg/hostlist"
mp "github.com/ClusterCockpit/cc-metric-collector/pkg/messageProcessor"
)
type IPMIReceiverClientConfig struct {
// Hostname the IPMI service belongs to
Protocol string // Protocol / tool to use for IPMI sensor reading
DriverType string // Out of band IPMI driver
Fanout int // Maximum number of simultaneous IPMI connections
NumHosts int // Number of remote IPMI devices with the same configuration
IPMIHosts string // List of remote IPMI devices to communicate with
IPMI2HostMapping map[string]string // Mapping between IPMI device name and host name
Username string // User name to authenticate with
Password string // Password to use for authentication
CLIOptions []string // Additional command line options for ipmi-sensors
isExcluded map[string]bool // is metric excluded
mp mp.MessageProcessor
}
type IPMIReceiver struct {
receiver
config struct {
defaultReceiverConfig
Interval time.Duration
// Client config for each IPMI hosts
ClientConfigs []IPMIReceiverClientConfig
}
// Storage for static information
//meta map[string]string
done chan bool // channel to finish / stop IPMI receiver
wg sync.WaitGroup // wait group for IPMI receiver
mp mp.MessageProcessor
}
// doReadMetrics reads metrics from all configure IPMI hosts.
func (r *IPMIReceiver) doReadMetric() {
for i := range r.config.ClientConfigs {
clientConfig := &r.config.ClientConfigs[i]
var cmd_options []string
if clientConfig.Protocol == "ipmi-sensors" {
cmd_options = append(cmd_options,
"--always-prefix",
"--sdr-cache-recreate",
// Attempt to interpret OEM data, such as event data, sensor readings, or general extra info
"--interpret-oem-data",
// Ignore not-available (i.e. N/A) sensors in output
"--ignore-not-available-sensors",
// Ignore unrecognized sensor events
"--ignore-unrecognized-events",
// Output fields in comma separated format
"--comma-separated-output",
// Do not output column headers
"--no-header-output",
// Output non-abbreviated units (e.g. 'Amps' instead of 'A').
// May aid in disambiguation of units (e.g. 'C' for Celsius or Coulombs).
"--non-abbreviated-units",
"--fanout", fmt.Sprint(clientConfig.Fanout),
"--driver-type", clientConfig.DriverType,
"--hostname", clientConfig.IPMIHosts,
"--username", clientConfig.Username,
"--password", clientConfig.Password,
)
cmd_options := append(cmd_options, clientConfig.CLIOptions...)
command := exec.Command("ipmi-sensors", cmd_options...)
stdout, _ := command.StdoutPipe()
errBuf := new(bytes.Buffer)
command.Stderr = errBuf
// start command
if err := command.Start(); err != nil {
cclog.ComponentError(
r.name,
fmt.Sprintf("doReadMetric(): Failed to start command \"%s\": %v", command.String(), err),
)
continue
}
// Read command output
const (
idxID = iota
idxName
idxType
idxReading
idxUnits
idxEvent
)
numPrefixRegex := regexp.MustCompile("^[[:digit:]][[:digit:]]-(.*)$")
scanner := bufio.NewScanner(stdout)
for scanner.Scan() {
// Read host
v1 := strings.Split(scanner.Text(), ": ")
if len(v1) != 2 {
continue
}
host, ok := clientConfig.IPMI2HostMapping[v1[0]]
if !ok {
continue
}
// Read sensors
v2 := strings.Split(v1[1], ",")
if len(v2) != 6 {
continue
}
// Skip sensors with non available sensor readings
if v2[idxReading] == "N/A" {
continue
}
metric := strings.ToLower(v2[idxType])
name := strings.ToLower(
strings.Replace(
strings.TrimSpace(
v2[idxName]), " ", "_", -1))
// remove prefix enumeration like 01-...
if v := numPrefixRegex.FindStringSubmatch(name); v != nil {
name = v[1]
}
unit := v2[idxUnits]
if unit == "Watts" {
// Power
metric = "power"
name = strings.TrimSuffix(name, "_power")
name = strings.TrimSuffix(name, "_pwr")
name = strings.TrimPrefix(name, "pwr_")
} else if metric == "voltage" &&
unit == "Volts" {
// Voltage
name = strings.TrimPrefix(name, "volt_")
} else if metric == "current" &&
unit == "Amps" {
// Current
unit = "Ampere"
} else if metric == "temperature" &&
unit == "degrees C" {
// Temperature
name = strings.TrimSuffix(name, "_temp")
unit = "degC"
} else if metric == "temperature" &&
unit == "degrees F" {
// Temperature
name = strings.TrimSuffix(name, "_temp")
unit = "degF"
} else if metric == "fan" && unit == "RPM" {
// Fan speed
metric = "fan_speed"
name = strings.TrimSuffix(name, "_tach")
name = strings.TrimPrefix(name, "spd_")
} else if (metric == "cooling device" ||
metric == "other units based sensor") &&
name == "system_air_flow" &&
unit == "CFM" {
// Air flow
metric = "air_flow"
name = strings.TrimSuffix(name, "_air_flow")
unit = "CubicFeetPerMinute"
} else if (metric == "processor" ||
metric == "other units based sensor") &&
(name == "cpu_utilization" ||
name == "io_utilization" ||
name == "mem_utilization" ||
name == "sys_utilization") &&
(unit == "unspecified" ||
unit == "%") {
// Utilization
metric = "utilization"
name = strings.TrimSuffix(name, "_utilization")
unit = "percent"
} else {
if false {
// Debug output for unprocessed metrics
fmt.Printf(
"host: '%s', metric: '%s', name: '%s', unit: '%s'\n",
host, metric, name, unit)
}
continue
}
// Skip excluded metrics
if clientConfig.isExcluded[metric] {
continue
}
// Parse sensor value
value, err := strconv.ParseFloat(v2[idxReading], 64)
if err != nil {
continue
}
y, err := lp.NewMessage(
metric,
map[string]string{
"hostname": host,
"type": "node",
"name": name,
},
map[string]string{
"source": r.name,
"group": "IPMI",
"unit": unit,
},
map[string]interface{}{
"value": value,
},
time.Now())
if err == nil {
mc, err := clientConfig.mp.ProcessMessage(y)
if err == nil && mc != nil {
m, err := r.mp.ProcessMessage(mc)
if err == nil && m != nil {
r.sink <- m
}
}
}
}
// Wait for command end
if err := command.Wait(); err != nil {
errMsg, _ := io.ReadAll(errBuf)
cclog.ComponentError(
r.name,
fmt.Sprintf("doReadMetric(): Failed to wait for the end of command \"%s\": %v\n",
strings.Replace(command.String(), clientConfig.Password, "<PW>", -1), err),
fmt.Sprintf("doReadMetric(): command stderr: \"%s\"\n", string(errMsg)),
)
}
}
}
}
func (r *IPMIReceiver) Start() {
cclog.ComponentDebug(r.name, "START")
// Start IPMI receiver
r.wg.Add(1)
go func() {
defer r.wg.Done()
// Create ticker
ticker := time.NewTicker(r.config.Interval)
defer ticker.Stop()
for {
r.doReadMetric()
select {
case tickerTime := <-ticker.C:
// Check if we missed the ticker event
if since := time.Since(tickerTime); since > 5*time.Second {
cclog.ComponentInfo(r.name, "Missed ticker event for more then", since)
}
// process ticker event -> continue
continue
case <-r.done:
// process done event
return
}
}
}()
cclog.ComponentDebug(r.name, "STARTED")
}
// Close receiver: close network connection, close files, close libraries, ...
func (r *IPMIReceiver) Close() {
cclog.ComponentDebug(r.name, "CLOSE")
// Send the signal and wait
close(r.done)
r.wg.Wait()
cclog.ComponentDebug(r.name, "DONE")
}
// NewIPMIReceiver creates a new instance of the redfish receiver
// Initialize the receiver by giving it a name and reading in the config JSON
func NewIPMIReceiver(name string, config json.RawMessage) (Receiver, error) {
var err error
r := new(IPMIReceiver)
// Config options from config file
configJSON := struct {
defaultReceiverConfig
// How often the IPMI sensor metrics should be read and send to the sink (default: 30 s)
IntervalString string `json:"interval,omitempty"`
// Maximum number of simultaneous IPMI connections (default: 64)
Fanout int `json:"fanout,omitempty"`
// Out of band IPMI driver (default: LAN_2_0)
DriverType string `json:"driver_type,omitempty"`
// Default client username, password and endpoint
Username *string `json:"username"` // User name to authenticate with
Password *string `json:"password"` // Password to use for authentication
Endpoint *string `json:"endpoint"` // URL of the IPMI device
// Globally excluded metrics
ExcludeMetrics []string `json:"exclude_metrics,omitempty"`
ClientConfigs []struct {
Fanout int `json:"fanout,omitempty"` // Maximum number of simultaneous IPMI connections (default: 64)
DriverType string `json:"driver_type,omitempty"` // Out of band IPMI driver (default: LAN_2_0)
HostList string `json:"host_list"` // List of hosts with the same client configuration
Username *string `json:"username"` // User name to authenticate with
Password *string `json:"password"` // Password to use for authentication
Endpoint *string `json:"endpoint"` // URL of the IPMI service
// Per client excluded metrics
ExcludeMetrics []string `json:"exclude_metrics,omitempty"`
// Additional command line options for ipmi-sensors
CLIOptions []string `json:"cli_options,omitempty"`
MessageProcessor json.RawMessage `json:"process_messages,omitempty"`
} `json:"client_config"`
}{
// Set defaults values
// Allow overwriting these defaults by reading config JSON
Fanout: 64,
DriverType: "LAN_2_0",
IntervalString: "30s",
}
// Set name of IPMIReceiver
r.name = fmt.Sprintf("IPMIReceiver(%s)", name)
// Create done channel
r.done = make(chan bool)
p, err := mp.NewMessageProcessor()
if err != nil {
return nil, fmt.Errorf("initialization of message processor failed: %v", err.Error())
}
r.mp = p
// Set static information
//r.meta = map[string]string{"source": r.name}
r.mp.AddAddMetaByCondition("true", "source", r.name)
// Read the IPMI receiver specific JSON config
if len(config) > 0 {
d := json.NewDecoder(bytes.NewReader(config))
d.DisallowUnknownFields()
if err := d.Decode(&configJSON); err != nil {
cclog.ComponentError(r.name, "Error reading config:", err.Error())
return nil, err
}
}
if len(r.config.MessageProcessor) > 0 {
err = r.mp.FromConfigJSON(r.config.MessageProcessor)
if err != nil {
return nil, fmt.Errorf("failed parsing JSON for message processor: %v", err.Error())
}
}
// Convert interval string representation to duration
r.config.Interval, err = time.ParseDuration(configJSON.IntervalString)
if err != nil {
err := fmt.Errorf(
"failed to parse duration string interval='%s': %w",
configJSON.IntervalString,
err,
)
cclog.Error(r.name, err)
return nil, err
}
// Create client config from JSON config
totalNumHosts := 0
for i := range configJSON.ClientConfigs {
clientConfigJSON := &configJSON.ClientConfigs[i]
var endpoint string
if clientConfigJSON.Endpoint != nil {
endpoint = *clientConfigJSON.Endpoint
} else if configJSON.Endpoint != nil {
endpoint = *configJSON.Endpoint
} else {
err := fmt.Errorf("client config number %v requires endpoint", i)
cclog.ComponentError(r.name, err)
return nil, err
}
fanout := configJSON.Fanout
if clientConfigJSON.Fanout != 0 {
fanout = clientConfigJSON.Fanout
}
driverType := configJSON.DriverType
if clientConfigJSON.DriverType != "" {
driverType = clientConfigJSON.DriverType
}
if driverType != "LAN" && driverType != "LAN_2_0" {
err := fmt.Errorf("client config number %v has invalid driver type %s", i, driverType)
cclog.ComponentError(r.name, err)
return nil, err
}
var protocol string
var host_pattern string
if e := strings.Split(endpoint, "://"); len(e) == 2 {
protocol = e[0]
host_pattern = e[1]
} else {
err := fmt.Errorf("client config number %v has invalid endpoint %s", i, endpoint)
cclog.ComponentError(r.name, err)
return nil, err
}
var username string
if clientConfigJSON.Username != nil {
username = *clientConfigJSON.Username
} else if configJSON.Username != nil {
username = *configJSON.Username
} else {
err := fmt.Errorf("client config number %v requires username", i)
cclog.ComponentError(r.name, err)
return nil, err
}
var password string
if clientConfigJSON.Password != nil {
password = *clientConfigJSON.Password
} else if configJSON.Password != nil {
password = *configJSON.Password
} else {
err := fmt.Errorf("client config number %v requires password", i)
cclog.ComponentError(r.name, err)
return nil, err
}
// Create mapping between IPMI host name and node host name
// This also guaranties that all IPMI host names are unique
ipmi2HostMapping := make(map[string]string)
hostList, err := hostlist.Expand(clientConfigJSON.HostList)
if err != nil {
err := fmt.Errorf("client config number %d failed to parse host list %s: %v",
i, clientConfigJSON.HostList, err)
cclog.ComponentError(r.name, err)
return nil, err
}
for _, host := range hostList {
ipmiHost := strings.Replace(host_pattern, "%h", host, -1)
ipmi2HostMapping[ipmiHost] = host
}
numHosts := len(ipmi2HostMapping)
totalNumHosts += numHosts
ipmiHostList := make([]string, 0, numHosts)
for ipmiHost := range ipmi2HostMapping {
ipmiHostList = append(ipmiHostList, ipmiHost)
}
// Additional command line options
for _, v := range clientConfigJSON.CLIOptions {
switch {
case v == "-u" || strings.HasPrefix(v, "--username"):
err := fmt.Errorf("client config number %v: do not set username in cli_options. Use json config username instead", i)
cclog.ComponentError(r.name, err)
return nil, err
case v == "-p" || strings.HasPrefix(v, "--password"):
err := fmt.Errorf("client config number %v: do not set password in cli_options. Use json config password instead", i)
cclog.ComponentError(r.name, err)
return nil, err
case v == "-h" || strings.HasPrefix(v, "--hostname"):
err := fmt.Errorf("client config number %v: do not set hostname in cli_options. Use json config host_list instead", i)
cclog.ComponentError(r.name, err)
return nil, err
case v == "-D" || strings.HasPrefix(v, "--driver-type"):
err := fmt.Errorf("client config number %v: do not set driver type in cli_options. Use json config driver_type instead", i)
cclog.ComponentError(r.name, err)
return nil, err
case v == "-F" || strings.HasPrefix(v, " --fanout"):
err := fmt.Errorf("client config number %v: do not set fanout in cli_options. Use json config fanout instead", i)
cclog.ComponentError(r.name, err)
return nil, err
case v == "--always-prefix" ||
v == "--sdr-cache-recreate" ||
v == "--interpret-oem-data" ||
v == "--ignore-not-available-sensors" ||
v == "--ignore-unrecognized-events" ||
v == "--comma-separated-output" ||
v == "--no-header-output" ||
v == "--non-abbreviated-units":
err := fmt.Errorf("client config number %v: Do not use option %s in cli_options, it is used internally", i, v)
cclog.ComponentError(r.name, err)
return nil, err
}
}
cliOptions := make([]string, 0)
cliOptions = append(cliOptions, clientConfigJSON.CLIOptions...)
// Is metrics excluded globally or per client
isExcluded := make(map[string]bool)
for _, key := range clientConfigJSON.ExcludeMetrics {
isExcluded[key] = true
}
for _, key := range configJSON.ExcludeMetrics {
isExcluded[key] = true
}
p, err := mp.NewMessageProcessor()
if err != nil {
return nil, fmt.Errorf("initialization of message processor failed: %v", err.Error())
}
if len(clientConfigJSON.MessageProcessor) > 0 {
err = p.FromConfigJSON(clientConfigJSON.MessageProcessor)
if err != nil {
return nil, fmt.Errorf("failed parsing JSON for message processor: %v", err.Error())
}
}
r.config.ClientConfigs = append(
r.config.ClientConfigs,
IPMIReceiverClientConfig{
Protocol: protocol,
Fanout: fanout,
DriverType: driverType,
NumHosts: numHosts,
IPMIHosts: strings.Join(ipmiHostList, ","),
IPMI2HostMapping: ipmi2HostMapping,
Username: username,
Password: password,
CLIOptions: cliOptions,
isExcluded: isExcluded,
mp: p,
})
}
if totalNumHosts == 0 {
err := fmt.Errorf("at least one IPMI host config is required")
cclog.ComponentError(r.name, err)
return nil, err
}
cclog.ComponentInfo(r.name, "monitoring", totalNumHosts, "IPMI hosts")
return r, nil
}

48
receivers/ipmiReceiver.md Normal file
View File

@@ -0,0 +1,48 @@
## IPMI Receiver
The IPMI Receiver uses `ipmi-sensors` from the [FreeIPMI](https://www.gnu.org/software/freeipmi/) project to read IPMI sensor readings and sensor data repository (SDR) information. The available metrics depend on the sensors provided by the hardware vendor but typically contain temperature, fan speed, voltage and power metrics.
### Configuration structure
```json
{
"<IPMI receiver name>": {
"type": "ipmi",
"interval": "30s",
"fanout": 256,
"username": "<Username>",
"password": "<Password>",
"endpoint": "ipmi-sensors://%h-bmc",
"exclude_metrics": [ "fan_speed", "voltage" ],
"client_config": [
{
"host_list": "n[1,2-4]"
},
{
"host_list": "n[5-6]",
"driver_type": "LAN",
"cli_options": [ "--workaround-flags=..." ],
"password": "<Password 2>"
}
]
}
}
```
Global settings:
- `interval`: How often the IPMI sensor metrics should be read and send to the sink (default: 30 s)
Global and per IPMI device settings (per IPMI device settings overwrite the global settings):
- `exclude_metrics`: list of excluded metrics e.g. fan_speed, power, temperature, utilization, voltage
- `fanout`: Maximum number of simultaneous IPMI connections (default: 64)
- `driver_type`: Out of band IPMI driver (default: LAN_2_0)
- `username`: User name to authenticate with
- `password`: Password to use for authentication
- `endpoint`: URL of the IPMI device (placeholder `%h` gets replaced by the hostname)
Per IPMI device settings:
- `host_list`: List of hosts with the same client configuration
- `cli_options`: Additional command line options for ipmi-sensors

View File

@@ -0,0 +1,45 @@
package receivers
import (
"encoding/json"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
mp "github.com/ClusterCockpit/cc-metric-collector/pkg/messageProcessor"
)
type defaultReceiverConfig struct {
Type string `json:"type"`
MessageProcessor json.RawMessage `json:"process_messages,omitempty"`
}
// Receiver configuration: Listen address, port
type ReceiverConfig struct {
Addr string `json:"address"`
Port string `json:"port"`
Database string `json:"database"`
Organization string `json:"organization,omitempty"`
Type string `json:"type"`
}
type receiver struct {
name string
sink chan lp.CCMessage
mp mp.MessageProcessor
}
type Receiver interface {
Start()
Close() // Close / finish metric receiver
Name() string // Name of the metric receiver
SetSink(sink chan lp.CCMessage) // Set sink channel
}
// Name returns the name of the metric receiver
func (r *receiver) Name() string {
return r.name
}
// SetSink set the sink channel
func (r *receiver) SetSink(sink chan lp.CCMessage) {
r.sink = sink
}

127
receivers/natsReceiver.go Normal file
View File

@@ -0,0 +1,127 @@
package receivers
import (
"encoding/json"
"errors"
"fmt"
"os"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
cclog "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
mp "github.com/ClusterCockpit/cc-metric-collector/pkg/messageProcessor"
nats "github.com/nats-io/nats.go"
)
type NatsReceiverConfig struct {
defaultReceiverConfig
Addr string `json:"address"`
Port string `json:"port"`
Subject string `json:"subject"`
User string `json:"user,omitempty"`
Password string `json:"password,omitempty"`
NkeyFile string `json:"nkey_file,omitempty"`
}
type NatsReceiver struct {
receiver
nc *nats.Conn
//meta map[string]string
config NatsReceiverConfig
}
// Start subscribes to the configured NATS subject
// Messages wil be handled by r._NatsReceive
func (r *NatsReceiver) Start() {
cclog.ComponentDebug(r.name, "START")
r.nc.Subscribe(r.config.Subject, r._NatsReceive)
}
// _NatsReceive receives subscribed messages from the NATS server
func (r *NatsReceiver) _NatsReceive(m *nats.Msg) {
if r.sink != nil {
messages, err := lp.FromBytes(m.Data)
if err != nil {
msg := "_NatsReceive: Failed to decode messages: " + err.Error()
cclog.ComponentError(r.name, msg)
}
for _, y := range messages {
m, err := r.mp.ProcessMessage(y)
if err == nil && m != nil && r.sink != nil {
r.sink <- m
}
}
}
}
// Close closes the connection to the NATS server
func (r *NatsReceiver) Close() {
if r.nc != nil {
cclog.ComponentDebug(r.name, "CLOSE")
r.nc.Close()
}
}
// NewNatsReceiver creates a new Receiver which subscribes to messages from a NATS server
func NewNatsReceiver(name string, config json.RawMessage) (Receiver, error) {
var uinfo nats.Option = nil
r := new(NatsReceiver)
r.name = fmt.Sprintf("NatsReceiver(%s)", name)
// Read configuration file, allow overwriting default config
r.config.Addr = "localhost"
r.config.Port = "4222"
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
}
}
if len(r.config.Addr) == 0 ||
len(r.config.Port) == 0 ||
len(r.config.Subject) == 0 {
return nil, errors.New("not all configuration variables set required by NatsReceiver")
}
p, err := mp.NewMessageProcessor()
if err != nil {
return nil, fmt.Errorf("initialization of message processor failed: %v", err.Error())
}
r.mp = p
if len(r.config.MessageProcessor) > 0 {
err = r.mp.FromConfigJSON(r.config.MessageProcessor)
if err != nil {
return nil, fmt.Errorf("failed parsing JSON for message processor: %v", err.Error())
}
}
// Set metadata
// r.meta = map[string]string{
// "source": r.name,
// }
r.mp.AddAddMetaByCondition("true", "source", r.name)
if len(r.config.User) > 0 && len(r.config.Password) > 0 {
uinfo = nats.UserInfo(r.config.User, r.config.Password)
} else if len(r.config.NkeyFile) > 0 {
_, err := os.Stat(r.config.NkeyFile)
if err == nil {
uinfo = nats.UserCredentials(r.config.NkeyFile)
} else {
cclog.ComponentError(r.name, "NKEY file", r.config.NkeyFile, "does not exist: %v", err.Error())
return nil, err
}
}
// Connect to NATS server
url := fmt.Sprintf("nats://%s:%s", r.config.Addr, r.config.Port)
cclog.ComponentDebug(r.name, "NewNatsReceiver", url, "Subject", r.config.Subject)
if nc, err := nats.Connect(url, uinfo); err == nil {
r.nc = nc
} else {
r.nc = nil
return nil, err
}
return r, nil
}

56
receivers/natsReceiver.md Normal file
View File

@@ -0,0 +1,56 @@
## `nats` receiver
The `nats` receiver can be used receive metrics from the NATS network. The `nats` receiver subscribes to the topic `database` and listens on `address` and `port` for metrics in the InfluxDB line protocol.
### Configuration structure
```json
{
"<name>": {
"type": "nats",
"address" : "nats-server.example.org",
"port" : "4222",
"subject" : "subject",
"user": "natsuser",
"password": "natssecret",
"nkey_file": "/path/to/nkey_file"
}
}
```
- `type`: makes the receiver a `nats` receiver
- `address`: Address of the NATS control server
- `port`: Port of the NATS control server
- `subject`: Subscribes to this subject and receive metrics
- `user`: Connect to nats using this user
- `password`: Connect to nats using this password
- `nkey_file`: Path to credentials file with NKEY
### Debugging
- Install NATS server and command line client
- Start NATS server
```bash
nats-server --net nats-server.example.org --port 4222
```
- Check NATS server works as expected
```bash
nats --server=nats-server-db.example.org:4222 server check
```
- Use NATS command line client to subscribe to all messages
```bash
nats --server=nats-server-db.example.org:4222 sub ">"
```
- Use NATS command line client to send message to NATS receiver
```bash
nats --server=nats-server-db.example.org:4222 pub subject \
"myMetric,hostname=myHost,type=hwthread,type-id=0,unit=Hz value=400000i 1694777161164284635
myMetric,hostname=myHost,type=hwthread,type-id=1,unit=Hz value=400001i 1694777161164284635"
```

View File

@@ -0,0 +1,122 @@
package receivers
import (
"bufio"
"encoding/json"
"errors"
"fmt"
"log"
"net/http"
"strconv"
"strings"
"sync"
"time"
cclog "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
)
type PrometheusReceiverConfig struct {
defaultReceiverConfig
Addr string `json:"address"`
Port string `json:"port"`
Path string `json:"path"`
Interval string `json:"interval"`
SSL bool `json:"ssl"`
}
type PrometheusReceiver struct {
receiver
meta map[string]string
config PrometheusReceiverConfig
interval time.Duration
done chan bool
wg sync.WaitGroup
ticker *time.Ticker
uri string
}
func (r *PrometheusReceiver) Start() {
cclog.ComponentDebug(r.name, "START", r.uri)
r.wg.Add(1)
r.ticker = time.NewTicker(r.interval)
go func() {
for {
select {
case <-r.done:
r.wg.Done()
return
case t := <-r.ticker.C:
resp, err := http.Get(r.uri)
if err != nil {
log.Fatal(err)
}
defer resp.Body.Close()
scanner := bufio.NewScanner(resp.Body)
for scanner.Scan() {
line := scanner.Text()
if strings.HasPrefix(line, "#") {
continue
}
lineSplit := strings.Fields(line)
// separate metric name from tags (labels in Prometheus)
tags := map[string]string{}
name := lineSplit[0]
if sindex := strings.Index(name, "{"); sindex >= 0 {
eindex := strings.Index(name, "}")
for _, kv := range strings.Split(name[sindex+1:eindex], ",") {
eq := strings.Index(kv, "=")
tags[kv[0:eq]] = strings.Trim(kv[eq+1:], "\"")
}
name = lineSplit[0][0:sindex]
}
value, err := strconv.ParseFloat(lineSplit[1], 64)
if err == nil {
y, err := lp.NewMessage(name, tags, r.meta, map[string]interface{}{"value": value}, t)
if err == nil {
r.sink <- y
}
}
}
}
}
}()
}
func (r *PrometheusReceiver) Close() {
cclog.ComponentDebug(r.name, "CLOSE")
r.done <- true
r.wg.Wait()
}
func NewPrometheusReceiver(name string, config json.RawMessage) (Receiver, error) {
r := new(PrometheusReceiver)
r.name = fmt.Sprintf("PrometheusReceiver(%s)", name)
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
}
}
if len(r.config.Addr) == 0 ||
len(r.config.Port) == 0 ||
len(r.config.Interval) == 0 {
return nil, errors.New("not all configuration variables set required by PrometheusReceiver (address and port)")
}
if len(r.config.Interval) > 0 {
t, err := time.ParseDuration(r.config.Interval)
if err == nil {
r.interval = t
}
}
r.meta = map[string]string{"source": r.name}
proto := "http"
if r.config.SSL {
proto = "https"
}
r.uri = fmt.Sprintf("%s://%s:%s/%s", proto, r.config.Addr, r.config.Port, r.config.Path)
return r, nil
}

View File

@@ -0,0 +1,27 @@
## `prometheus` receiver
The `prometheus` receiver can be used to scrape the metrics of a single `prometheus` client. It does **not** use any official Golang library but making simple HTTP get requests and parse the response.
### Configuration structure
```json
{
"<name>": {
"type": "prometheus",
"address" : "testpromhost",
"port" : "12345",
"path" : "/prometheus",
"interval": "5s",
"ssl" : true,
}
}
```
- `type`: makes the receiver a `prometheus` receiver
- `address`: Hostname or IP of the Prometheus agent
- `port`: Port of Prometheus agent
- `path`: Path to the Prometheus endpoint
- `interval`: Scrape the Prometheus endpoint in this interval (default '5s')
- `ssl`: Use SSL or not
The receiver requests data from `http(s)://<address>:<port>/<path>`.

122
receivers/receiveManager.go Normal file
View File

@@ -0,0 +1,122 @@
package receivers
import (
"encoding/json"
"fmt"
"os"
"sync"
cclog "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
)
var AvailableReceivers = map[string]func(name string, config json.RawMessage) (Receiver, error){
"http": NewHttpReceiver,
"ipmi": NewIPMIReceiver,
"nats": NewNatsReceiver,
"redfish": NewRedfishReceiver,
}
type receiveManager struct {
inputs []Receiver
output chan lp.CCMessage
config []json.RawMessage
}
type ReceiveManager interface {
Init(wg *sync.WaitGroup, receiverConfigFile string) error
AddInput(name string, rawConfig json.RawMessage) error
AddOutput(output chan lp.CCMessage)
Start()
Close()
}
func (rm *receiveManager) Init(wg *sync.WaitGroup, receiverConfigFile string) error {
// Initialize struct fields
rm.inputs = make([]Receiver, 0)
rm.output = nil
rm.config = make([]json.RawMessage, 0)
configFile, err := os.Open(receiverConfigFile)
if err != nil {
cclog.ComponentError("ReceiveManager", err.Error())
return err
}
defer configFile.Close()
jsonParser := json.NewDecoder(configFile)
var rawConfigs map[string]json.RawMessage
err = jsonParser.Decode(&rawConfigs)
if err != nil {
cclog.ComponentError("ReceiveManager", err.Error())
return err
}
for name, raw := range rawConfigs {
rm.AddInput(name, raw)
}
return nil
}
func (rm *receiveManager) Start() {
cclog.ComponentDebug("ReceiveManager", "START")
for _, r := range rm.inputs {
cclog.ComponentDebug("ReceiveManager", "START", r.Name())
r.Start()
}
cclog.ComponentDebug("ReceiveManager", "STARTED")
}
func (rm *receiveManager) AddInput(name string, rawConfig json.RawMessage) error {
var config defaultReceiverConfig
err := json.Unmarshal(rawConfig, &config)
if err != nil {
cclog.ComponentError("ReceiveManager", "SKIP", config.Type, "JSON config error:", err.Error())
return err
}
if config.Type == "" {
cclog.ComponentError("ReceiveManager", "SKIP", "JSON config for receiver", name, "does not contain a receiver type")
return fmt.Errorf("JSON config for receiver %s does not contain a receiver type", name)
}
if _, found := AvailableReceivers[config.Type]; !found {
cclog.ComponentError("ReceiveManager", "SKIP", "unknown receiver type:", config.Type)
return fmt.Errorf("unknown receiver type: %s", config.Type)
}
r, err := AvailableReceivers[config.Type](name, rawConfig)
if err != nil {
cclog.ComponentError("ReceiveManager", "SKIP", name, "initialization failed:", err.Error())
return err
}
rm.inputs = append(rm.inputs, r)
rm.config = append(rm.config, rawConfig)
cclog.ComponentDebug("ReceiveManager", "ADD RECEIVER", r.Name())
return nil
}
func (rm *receiveManager) AddOutput(output chan lp.CCMessage) {
rm.output = output
for _, r := range rm.inputs {
r.SetSink(rm.output)
}
}
func (rm *receiveManager) Close() {
cclog.ComponentDebug("ReceiveManager", "CLOSE")
// Close all receivers
for _, r := range rm.inputs {
cclog.ComponentDebug("ReceiveManager", "CLOSE", r.Name())
r.Close()
}
cclog.ComponentDebug("ReceiveManager", "DONE")
}
func New(wg *sync.WaitGroup, receiverConfigFile string) (ReceiveManager, error) {
r := new(receiveManager)
err := r.Init(wg, receiverConfigFile)
if err != nil {
return nil, err
}
return r, err
}

1055
receivers/redfishReceiver.go Normal file

File diff suppressed because it is too large Load Diff

View File

@@ -0,0 +1,65 @@
## Redfish receiver
The Redfish receiver uses the [Redfish (specification)](https://www.dmtf.org/standards/redfish) to query thermal and power metrics. Thermal metrics may include various fan speeds and temperatures. Power metrics may include the current power consumption of various hardware components. It may also include the minimum, maximum and average power consumption of these components in a given time interval. The receiver will poll each configured redfish device once in a given interval. Multiple devices can be accessed in parallel to increase throughput.
### Configuration structure
```json
{
"<redfish receiver name>": {
"type": "redfish",
"username": "<Username>",
"password": "<Password>",
"endpoint": "https://%h-bmc",
"exclude_metrics": [ "min_consumed_watts" ],
"client_config": [
{
"host_list": "n[1,2-4]"
},
{
"host_list": "n5",
"disable_power_metrics": true,
"disable_processor_metrics": true,
"disable_thermal_metrics": true
},
{
"host_list": "n6" ],
"username": "<Username 2>",
"password": "<Password 2>",
"endpoint": "https://%h-BMC",
"disable_sensor_metrics": true
}
]
}
}
```
Global settings:
- `fanout`: Maximum number of simultaneous redfish connections (default: 64)
- `interval`: How often the redfish power metrics should be read and send to the sink (default: 30 s)
- `http_insecure`: Control whether a client verifies the server's certificate (default: true == do not verify server's certificate)
- `http_timeout`: Time limit for requests made by this HTTP client (default: 10 s)
Global and per redfish device settings (per redfish device settings overwrite the global settings):
- `disable_power_metrics`:
disable collection of power metrics
(`/redfish/v1/Chassis/{ChassisId}/Power`)
- `disable_processor_metrics`:
disable collection of processor metrics
(`/redfish/v1/Systems/{ComputerSystemId}/Processors/{ProcessorId}/ProcessorMetrics`)
- `disable_sensors`:
disable collection of fan, power and thermal sensor metrics
(`/redfish/v1/Chassis/{ChassisId}/Sensors/{SensorId}`)
- `disable_thermal_metrics`:
disable collection of thermal metrics
(`/redfish/v1/Chassis/{ChassisId}/Thermal`)
- `exclude_metrics`: list of excluded metrics
- `username`: User name to authenticate with
- `password`: Password to use for authentication
- `endpoint`: URL of the redfish service (placeholder `%h` gets replaced by the hostname)
Per redfish device settings:
- `host_list`: List of hosts with the same client configuration

122
receivers/sampleReceiver.go Normal file
View File

@@ -0,0 +1,122 @@
package receivers
import (
"encoding/json"
"fmt"
cclog "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
mp "github.com/ClusterCockpit/cc-metric-collector/pkg/messageProcessor"
)
// SampleReceiver configuration: receiver type, listen address, port
// The defaultReceiverConfig contains the keys 'type' and 'process_messages'
type SampleReceiverConfig struct {
defaultReceiverConfig
Addr string `json:"address"`
Port string `json:"port"`
}
type SampleReceiver struct {
receiver
config SampleReceiverConfig
// Storage for static information
// Use in case of own go routine
// done chan bool
// wg sync.WaitGroup
}
// Implement functions required for Receiver interface
// Start(), Close()
// See: metricReceiver.go
func (r *SampleReceiver) Start() {
cclog.ComponentDebug(r.name, "START")
// Start server process like http.ListenAndServe()
// 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() {
// 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
// }
// }
// }()
}
// Close receiver: close network connection, close files, close libraries, ...
func (r *SampleReceiver) Close() {
cclog.ComponentDebug(r.name, "CLOSE")
// Close server like http.Shutdown()
// in case of own go routine, send the signal and wait
// r.done <- true
// r.wg.Wait()
}
// 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 NewSampleReceiver(name string, config json.RawMessage) (Receiver, error) {
r := new(SampleReceiver)
// Set name of SampleReceiver
// The name should be chosen in such a way that different instances of SampleReceiver can be distinguished
r.name = fmt.Sprintf("SampleReceiver(%s)", name)
// create new message processor
p, err := mp.NewMessageProcessor()
if err != nil {
cclog.ComponentError(r.name, "Initialization of message processor failed:", err.Error())
return nil, fmt.Errorf("initialization of message processor failed: %v", err.Error())
}
r.mp = p
// Set static information
err = r.mp.AddAddMetaByCondition("true", "source", r.name)
if err != nil {
cclog.ComponentError(r.name, fmt.Sprintf("Failed to add static information source=%s:", r.name), err.Error())
return nil, fmt.Errorf("failed to add static information source=%s: %v", r.name, err.Error())
}
// Set defaults in r.config
// Allow overwriting these defaults by reading config JSON
// Read the sample 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
}
}
// Add message processor config
if len(r.config.MessageProcessor) > 0 {
err = r.mp.FromConfigJSON(r.config.MessageProcessor)
if err != nil {
cclog.ComponentError(r.name, "Failed parsing JSON for message processor:", err.Error())
return nil, fmt.Errorf("failed parsing JSON for message processor: %v", err.Error())
}
}
// Check that all required fields in the configuration are set
// Use 'if len(r.config.Option) > 0' for strings
return r, nil
}

108
sinks/README.md Normal file
View File

@@ -0,0 +1,108 @@
# CCMetric sinks
This folder contains the SinkManager and sink implementations for the cc-metric-collector.
# Available sinks:
- [`stdout`](./stdoutSink.md): Print all metrics to `stdout`, `stderr` or a file
- [`http`](./httpSink.md): Send metrics to an HTTP server as POST requests
- [`influxdb`](./influxSink.md): Send metrics to an [InfluxDB](https://www.influxdata.com/products/influxdb/) database
- [`influxasync`](./influxAsyncSink.md): Send metrics to an [InfluxDB](https://www.influxdata.com/products/influxdb/) database with non-blocking write API
- [`nats`](./natsSink.md): Publish metrics to the [NATS](https://nats.io/) network overlay system
- [`ganglia`](./gangliaSink.md): Publish metrics in the [Ganglia Monitoring System](http://ganglia.info/) using the `gmetric` CLI tool
- [`libganglia`](./libgangliaSink.md): Publish metrics in the [Ganglia Monitoring System](http://ganglia.info/) directly using `libganglia.so`
- [`prometeus`](./prometheusSink.md): Publish metrics for the [Prometheus Monitoring System](https://prometheus.io/)
# Configuration
The configuration file for the sinks is a list of configurations. The `type` field in each specifies which sink to initialize.
```json
{
"mystdout" : {
"type" : "stdout",
"meta_as_tags" : [
"unit"
]
},
"metricstore" : {
"type" : "http",
"host" : "localhost",
"port" : "4123",
"database" : "ccmetric",
"password" : "<jwt token>"
}
}
```
# Contributing own sinks
A sink contains five functions and is derived from the type `sink`:
* `Init(name string, config json.RawMessage) error`
* `Write(point CCMetric) error`
* `Flush() error`
* `Close()`
* `New<Typename>(name string, config json.RawMessage) (Sink, error)` (calls the `Init()` function)
The data structures should be set up in `Init()` like opening a file or server connection. The `Write()` function writes/sends the data. For non-blocking sinks, the `Flush()` method tells the sink to drain its internal buffers. The `Close()` function should tear down anything created in `Init()`.
Finally, the sink needs to be registered in the `sinkManager.go`. There is a list of sinks called `AvailableSinks` which is a map (`sink_type_string` -> `pointer to sink interface`). Add a new entry with a descriptive name and the new sink.
## Sample sink
```go
package sinks
import (
"encoding/json"
"log"
lp "github.com/ClusterCockpit/cc-metric-collector/internal/ccMetric"
)
type SampleSinkConfig struct {
defaultSinkConfig // defines JSON tags for 'name' and 'meta_as_tags'
}
type SampleSink struct {
sink // declarate 'name' and 'meta_as_tags'
config StdoutSinkConfig // entry point to the SampleSinkConfig
}
// Initialize the sink by giving it a name and reading in the config JSON
func (s *SampleSink) Init(name string, config json.RawMessage) error {
s.name = fmt.Sprintf("SampleSink(%s)", name) // Always specify a name here
// Read in the config JSON
if len(config) > 0 {
err := json.Unmarshal(config, &s.config)
if err != nil {
return err
}
}
return nil
}
// Code to submit a single CCMetric to the sink
func (s *SampleSink) Write(point lp.CCMetric) error {
log.Print(point)
return nil
}
// If the sink uses batched sends internally, you can tell to flush its buffers
func (s *SampleSink) Flush() error {
return nil
}
// Close sink: close network connection, close files, close libraries, ...
func (s *SampleSink) Close() {}
// New function to create a new instance of the sink
func NewSampleSink(name string, config json.RawMessage) (Sink, error) {
s := new(SampleSink)
err := s.Init(name, config)
return s, err
}
```

269
sinks/gangliaCommon.go Normal file
View File

@@ -0,0 +1,269 @@
package sinks
import (
"fmt"
"strings"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
)
func GangliaMetricName(point lp.CCMessage) string {
name := point.Name()
metricType, typeOK := point.GetTag("type")
metricTid, tidOk := point.GetTag("type-id")
gangliaType := metricType + metricTid
if strings.Contains(name, metricType) && tidOk {
name = strings.Replace(name, metricType, gangliaType, -1)
} else if typeOK && tidOk {
name = metricType + metricTid + "_" + name
} else if point.HasTag("device") {
device, _ := point.GetTag("device")
name = name + "_" + device
}
return name
}
func GangliaMetricRename(name string) string {
if name == "net_bytes_in" {
return "bytes_in"
} else if name == "net_bytes_out" {
return "bytes_out"
} else if name == "net_pkts_in" {
return "pkts_in"
} else if name == "net_pkts_out" {
return "pkts_out"
} else if name == "cpu_iowait" {
return "cpu_wio"
}
return name
}
func GangliaSlopeType(point lp.CCMessage) uint {
name := point.Name()
if name == "mem_total" || name == "swap_total" {
return 0
}
return 3
}
const DEFAULT_GANGLIA_METRIC_TMAX = 300
const DEFAULT_GANGLIA_METRIC_SLOPE = "both"
type GangliaMetric struct {
Name string
Type string
Slope string
Tmax int
Unit string
}
type GangliaMetricGroup struct {
Name string
Metrics []GangliaMetric
}
var CommonGangliaMetrics = []GangliaMetricGroup{
{
Name: "memory",
Metrics: []GangliaMetric{
{"mem_total", "float", "zero", 1200, "KB"},
{"swap_total", "float", "zero", 1200, "KB"},
{"mem_free", "float", "both", 180, "KB"},
{"mem_shared", "float", "both", 180, "KB"},
{"mem_buffers", "float", "both", 180, "KB"},
{"mem_cached", "float", "both", 180, "KB"},
{"swap_free", "float", "both", 180, "KB"},
{"mem_sreclaimable", "float", "both", 180, "KB"},
{"mem_slab", "float", "both", 180, "KB"},
},
},
{
Name: "cpu",
Metrics: []GangliaMetric{
{"cpu_num", "uint32", "zero", 1200, "CPUs"},
{"cpu_speed", "uint32", "zero", 1200, "MHz"},
{"cpu_user", "float", "both", 90, "%"},
{"cpu_nice", "float", "both", 90, "%"},
{"cpu_system", "float", "both", 90, "%"},
{"cpu_idle", "float", "both", 3800, "%"},
{"cpu_aidle", "float", "both", 90, "%"},
{"cpu_wio", "float", "both", 90, "%"},
{"cpu_intr", "float", "both", 90, "%"},
{"cpu_sintr", "float", "both", 90, "%"},
{"cpu_steal", "float", "both", 90, "%"},
{"cpu_guest", "float", "both", 90, "%"},
{"cpu_gnice", "float", "both", 90, "%"},
},
},
{
Name: "load",
Metrics: []GangliaMetric{
{"load_one", "float", "both", 70, ""},
{"load_five", "float", "both", 325, ""},
{"load_fifteen", "float", "both", 950, ""},
},
},
{
Name: "disk",
Metrics: []GangliaMetric{
{"disk_total", "double", "both", 1200, "GB"},
{"disk_free", "double", "both", 180, "GB"},
{"part_max_used", "float", "both", 180, "%"},
},
},
{
Name: "network",
Metrics: []GangliaMetric{
{"bytes_out", "float", "both", 300, "bytes/sec"},
{"bytes_in", "float", "both", 300, "bytes/sec"},
{"pkts_in", "float", "both", 300, "packets/sec"},
{"pkts_out", "float", "both", 300, "packets/sec"},
},
},
{
Name: "process",
Metrics: []GangliaMetric{
{"proc_run", "uint32", "both", 950, ""},
{"proc_total", "uint32", "both", 950, ""},
},
},
{
Name: "system",
Metrics: []GangliaMetric{
{"boottime", "uint32", "zero", 1200, "s"},
{"sys_clock", "uint32", "zero", 1200, "s"},
{"machine_type", "string", "zero", 1200, ""},
{"os_name", "string", "zero", 1200, ""},
{"os_release", "string", "zero", 1200, ""},
{"mtu", "uint32", "both", 1200, ""},
},
},
}
type GangliaMetricConfig struct {
Type string
Slope string
Tmax int
Unit string
Group string
Value string
Name string
}
func GetCommonGangliaConfig(point lp.CCMessage) 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 {
valueStr := ""
value, ok := point.GetField("value")
if ok {
switch real := value.(type) {
case float64:
valueStr = fmt.Sprintf("%f", real)
case float32:
valueStr = fmt.Sprintf("%f", real)
case int64:
valueStr = fmt.Sprintf("%d", real)
case int32:
valueStr = fmt.Sprintf("%d", real)
case int:
valueStr = fmt.Sprintf("%d", real)
case uint64:
valueStr = fmt.Sprintf("%d", real)
case uint32:
valueStr = fmt.Sprintf("%d", real)
case uint:
valueStr = fmt.Sprintf("%d", real)
case string:
valueStr = real
default:
}
}
return GangliaMetricConfig{
Group: group.Name,
Type: metric.Type,
Slope: metric.Slope,
Tmax: metric.Tmax,
Unit: metric.Unit,
Value: valueStr,
Name: GangliaMetricRename(mname),
}
}
}
}
return GangliaMetricConfig{
Group: "",
Type: "",
Slope: "",
Tmax: 0,
Unit: "",
Value: "",
Name: "",
}
}
func GetGangliaConfig(point lp.CCMessage) 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
}
unit := ""
if u, ok := point.GetMeta("unit"); ok {
unit = u
}
valueType := "double"
valueStr := ""
value, ok := point.GetField("value")
if ok {
switch real := value.(type) {
case float64:
valueStr = fmt.Sprintf("%f", real)
valueType = "double"
case float32:
valueStr = fmt.Sprintf("%f", real)
valueType = "float"
case int64:
valueStr = fmt.Sprintf("%d", real)
valueType = "int32"
case int32:
valueStr = fmt.Sprintf("%d", real)
valueType = "int32"
case int:
valueStr = fmt.Sprintf("%d", real)
valueType = "int32"
case uint64:
valueStr = fmt.Sprintf("%d", real)
valueType = "uint32"
case uint32:
valueStr = fmt.Sprintf("%d", real)
valueType = "uint32"
case uint:
valueStr = fmt.Sprintf("%d", real)
valueType = "uint32"
case string:
valueStr = real
valueType = "string"
default:
valueType = "invalid"
}
}
return GangliaMetricConfig{
Group: group,
Type: valueType,
Slope: DEFAULT_GANGLIA_METRIC_SLOPE,
Tmax: DEFAULT_GANGLIA_METRIC_TMAX,
Unit: unit,
Value: valueStr,
Name: GangliaMetricRename(mname),
}
}

147
sinks/gangliaSink.go Normal file
View File

@@ -0,0 +1,147 @@
package sinks
import (
"bytes"
"encoding/json"
"errors"
"fmt"
"strings"
// "time"
"os/exec"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
cclog "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
mp "github.com/ClusterCockpit/cc-metric-collector/pkg/messageProcessor"
)
const GMETRIC_EXEC = `gmetric`
const GMETRIC_CONFIG = `/etc/ganglia/gmond.conf`
type GangliaSinkConfig struct {
defaultSinkConfig
GmetricPath string `json:"gmetric_path,omitempty"`
GmetricConfig string `json:"gmetric_config,omitempty"`
AddGangliaGroup bool `json:"add_ganglia_group,omitempty"`
AddTagsAsDesc bool `json:"add_tags_as_desc,omitempty"`
ClusterName string `json:"cluster_name,omitempty"`
AddTypeToName bool `json:"add_type_to_name,omitempty"`
AddUnits bool `json:"add_units,omitempty"`
}
type GangliaSink struct {
sink
gmetric_path string
gmetric_config string
config GangliaSinkConfig
}
func (s *GangliaSink) Write(msg lp.CCMessage) error {
var err error = nil
//var tagsstr []string
var argstr []string
point, err := s.mp.ProcessMessage(msg)
if err == nil && point != nil {
// 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 %q (Ganglia name %q) has no 'value' field", point.Name(), conf.Name)
}
if s.config.AddGangliaGroup {
argstr = append(argstr, fmt.Sprintf("--group=%s", conf.Group))
}
if s.config.AddUnits && len(conf.Unit) > 0 {
argstr = append(argstr, fmt.Sprintf("--units=%s", conf.Unit))
}
if len(s.config.ClusterName) > 0 {
argstr = append(argstr, fmt.Sprintf("--cluster=%s", s.config.ClusterName))
}
// if s.config.AddTagsAsDesc && len(tagsstr) > 0 {
// argstr = append(argstr, fmt.Sprintf("--desc=%q", strings.Join(tagsstr, ",")))
// }
if len(s.gmetric_config) > 0 {
argstr = append(argstr, fmt.Sprintf("--conf=%s", s.gmetric_config))
}
if s.config.AddTypeToName {
argstr = append(argstr, fmt.Sprintf("--name=%s", GangliaMetricName(point)))
} else {
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))
argstr = append(argstr, fmt.Sprintf("--type=%s", conf.Type))
argstr = append(argstr, fmt.Sprintf("--tmax=%d", conf.Tmax))
cclog.ComponentDebug(s.name, s.gmetric_path, strings.Join(argstr, " "))
command := exec.Command(s.gmetric_path, argstr...)
command.Wait()
_, err = command.Output()
}
return err
}
func (s *GangliaSink) Flush() error {
return nil
}
func (s *GangliaSink) Close() {
}
func NewGangliaSink(name string, config json.RawMessage) (Sink, error) {
s := new(GangliaSink)
s.name = fmt.Sprintf("GangliaSink(%s)", name)
s.config.AddTagsAsDesc = false
s.config.AddGangliaGroup = false
if len(config) > 0 {
d := json.NewDecoder(bytes.NewReader(config))
d.DisallowUnknownFields()
if err := d.Decode(&s.config); err != nil {
cclog.ComponentError(s.name, "Error reading config:", err.Error())
return nil, err
}
}
s.gmetric_path = ""
s.gmetric_config = ""
p, err := mp.NewMessageProcessor()
if err != nil {
return nil, fmt.Errorf("initialization of message processor failed: %v", err.Error())
}
s.mp = p
if len(s.config.GmetricPath) > 0 {
p, err := exec.LookPath(s.config.GmetricPath)
if err == nil {
s.gmetric_path = p
}
}
if len(s.gmetric_path) == 0 {
p, err := exec.LookPath(string(GMETRIC_EXEC))
if err == nil {
s.gmetric_path = p
}
}
if len(s.gmetric_path) == 0 {
return nil, errors.New("cannot find executable 'gmetric'")
}
if len(s.config.GmetricConfig) > 0 {
s.gmetric_config = s.config.GmetricConfig
}
if len(s.config.MessageProcessor) > 0 {
err = s.mp.FromConfigJSON(s.config.MessageProcessor)
if err != nil {
return nil, fmt.Errorf("failed parsing JSON for message processor: %v", err.Error())
}
}
for _, k := range s.config.MetaAsTags {
s.mp.AddMoveMetaToTags("true", k, k)
}
return s, nil
}

25
sinks/gangliaSink.md Normal file
View File

@@ -0,0 +1,25 @@
## `ganglia` sink
The `ganglia` sink uses the `gmetric` tool of the [Ganglia Monitoring System](http://ganglia.info/) to submit the metrics
### Configuration structure
```json
{
"<name>": {
"type": "ganglia",
"gmetric_path" : "/path/to/gmetric",
"add_ganglia_group" : true,
"process_messages" : {
"see" : "docs of message processor for valid fields"
},
"meta_as_tags" : []
}
}
```
- `type`: makes the sink an `ganglia` sink
- `gmetric_path`: Path to `gmetric` executable (optional). If not given, the sink searches in `$PATH` for `gmetric`.
- `add_ganglia_group`: Add `--group=X` based on meta information to the `gmetric` call. Some old versions of `gmetric` do not support the `--group` option.
- `process_messages`: Process messages with given rules before progressing or dropping, see [here](../pkg/messageProcessor/README.md) (optional)
- `meta_as_tags`: print all meta information as tags in the output (deprecated, optional)

306
sinks/httpSink.go Normal file
View File

@@ -0,0 +1,306 @@
package sinks
import (
"bytes"
"encoding/json"
"errors"
"fmt"
"net/http"
"sync"
"time"
lp "github.com/ClusterCockpit/cc-energy-manager/pkg/cc-message"
cclog "github.com/ClusterCockpit/cc-metric-collector/pkg/ccLogger"
mp "github.com/ClusterCockpit/cc-metric-collector/pkg/messageProcessor"
influx "github.com/influxdata/line-protocol/v2/lineprotocol"
"golang.org/x/exp/slices"
)
type HttpSinkConfig struct {
defaultSinkConfig
// The full URL of the endpoint
URL string `json:"url"`
// JSON web tokens for authentication (Using the *Bearer* scheme)
JWT string `json:"jwt,omitempty"`
// Basic authentication
Username string `json:"username"`
Password string `json:"password"`
useBasicAuth bool
// time limit for requests made by the http client
Timeout string `json:"timeout,omitempty"`
timeout time.Duration
// Maximum amount of time an idle (keep-alive) connection will remain idle before closing itself
// should be larger than the measurement interval to keep the connection open
IdleConnTimeout string `json:"idle_connection_timeout,omitempty"`
idleConnTimeout time.Duration
// Batch all writes arriving in during this duration
// (default '5s', batching can be disabled by setting it to 0)
FlushDelay string `json:"flush_delay,omitempty"`
flushDelay time.Duration
// Maximum number of retries to connect to the http server (default: 3)
MaxRetries int `json:"max_retries,omitempty"`
// Timestamp precision
Precision string `json:"precision,omitempty"`
}
type HttpSink struct {
sink
client *http.Client
// influx line protocol encoder
encoder influx.Encoder
// Flush() runs in another goroutine and accesses the influx line protocol encoder,
// so this encoderLock has to protect the encoder
encoderLock sync.Mutex
// timer to run Flush()
flushTimer *time.Timer
// Lock to assure that only one timer is running at a time
timerLock sync.Mutex
config HttpSinkConfig
}
// Write sends metric m as http message
func (s *HttpSink) Write(msg lp.CCMessage) error {
// submit m only after applying processing/dropping rules
m, err := s.mp.ProcessMessage(msg)
if err == nil && m != nil {
// Lock for encoder usage
s.encoderLock.Lock()
err = EncoderAdd(&s.encoder, m)
// Unlock encoder usage
s.encoderLock.Unlock()
// Check that encoding worked
if err != nil {
return fmt.Errorf("encoding failed: %v", err)
}
}
if s.config.flushDelay == 0 {
// Directly flush if no flush delay is configured
return s.Flush()
} else if s.timerLock.TryLock() {
// Setup flush timer when flush delay is configured
// and no other timer is already running
if s.flushTimer != nil {
// Restarting existing flush timer
cclog.ComponentDebug(s.name, "Write(): Restarting flush timer")
s.flushTimer.Reset(s.config.flushDelay)
} else {
// Creating and starting flush timer
cclog.ComponentDebug(s.name, "Write(): Starting new flush timer")
s.flushTimer = time.AfterFunc(
s.config.flushDelay,
func() {
defer s.timerLock.Unlock()
cclog.ComponentDebug(s.name, "Starting flush triggered by flush timer")
if err := s.Flush(); err != nil {
cclog.ComponentError(s.name, "Flush triggered by flush timer: flush failed:", err)
}
})
}
}
return nil
}
// Flush sends all metrics stored in encoder to HTTP server
func (s *HttpSink) Flush() error {
// Lock for encoder usage
// Own lock for as short as possible: the time it takes to clone the buffer.
s.encoderLock.Lock()
buf := slices.Clone(s.encoder.Bytes())
s.encoder.Reset()
// Unlock encoder usage
s.encoderLock.Unlock()
if len(buf) == 0 {
return nil
}
cclog.ComponentDebug(s.name, "Flush(): Flushing")
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, "Flush(): Failed to create HTTP request:", err)
return err
}
// Set authorization header
if len(s.config.JWT) != 0 {
req.Header.Set("Authorization", fmt.Sprintf("Bearer %s", s.config.JWT))
}
// Set basic authentication
if s.config.useBasicAuth {
req.SetBasicAuth(s.config.Username, s.config.Password)
}
// Do request
res, err = s.client.Do(req)
if err != nil {
cclog.ComponentError(s.name, "Flush(): transport/tcp error:", err)
// Wait between retries
time.Sleep(time.Duration(i+1) * (time.Second / 2))
continue
}
break
}
if res == nil {
return errors.New("flush failed due to repeated errors")
}
// Handle application errors
if res.StatusCode != http.StatusOK {
err := errors.New(res.Status)
cclog.ComponentError(s.name, "Flush(): Application error:", err)
return err
}
return nil
}
func (s *HttpSink) Close() {
cclog.ComponentDebug(s.name, "Closing HTTP connection")
// Stop existing timer and immediately flush
if s.flushTimer != nil {
if ok := s.flushTimer.Stop(); ok {
s.timerLock.Unlock()
}
}
// Flush
if err := s.Flush(); err != nil {
cclog.ComponentError(s.name, "Close(): Flush failed:", err)
}
s.client.CloseIdleConnections()
}
// NewHttpSink creates a new http sink
func NewHttpSink(name string, config json.RawMessage) (Sink, error) {
s := new(HttpSink)
// Set default values
s.name = fmt.Sprintf("HttpSink(%s)", name)
// should be larger than the measurement interval to keep the connection open
s.config.IdleConnTimeout = "120s"
s.config.Timeout = "5s"
s.config.FlushDelay = "5s"
s.config.MaxRetries = 3
s.config.Precision = "s"
cclog.ComponentDebug(s.name, "Init()")
// Read config
if len(config) > 0 {
d := json.NewDecoder(bytes.NewReader(config))
d.DisallowUnknownFields()
if err := d.Decode(&s.config); err != nil {
cclog.ComponentError(s.name, "Error reading config:", err.Error())
return nil, err
}
}
if len(s.config.URL) == 0 {
return nil, errors.New("`url` config option is required for HTTP sink")
}
// Check basic authentication config
if len(s.config.Username) > 0 || len(s.config.Password) > 0 {
s.config.useBasicAuth = true
}
if s.config.useBasicAuth && len(s.config.Username) == 0 {
return nil, errors.New("basic authentication requires username")
}
if s.config.useBasicAuth && len(s.config.Password) == 0 {
return nil, errors.New("basic authentication requires password")
}
p, err := mp.NewMessageProcessor()
if err != nil {
return nil, fmt.Errorf("initialization of message processor failed: %v", err.Error())
}
s.mp = p
if len(s.config.IdleConnTimeout) > 0 {
t, err := time.ParseDuration(s.config.IdleConnTimeout)
if err == nil {
cclog.ComponentDebug(s.name, "Init(): idleConnTimeout", t)
s.config.idleConnTimeout = t
}
}
if len(s.config.Timeout) > 0 {
t, err := time.ParseDuration(s.config.Timeout)
if err == nil {
s.config.timeout = t
cclog.ComponentDebug(s.name, "Init(): timeout", t)
}
}
if len(s.config.FlushDelay) > 0 {
t, err := time.ParseDuration(s.config.FlushDelay)
if err == nil {
s.config.flushDelay = t
cclog.ComponentDebug(s.name, "Init(): flushDelay", t)
}
}
if len(s.config.MessageProcessor) > 0 {
err = p.FromConfigJSON(s.config.MessageProcessor)
if err != nil {
return nil, fmt.Errorf("failed parsing JSON for message processor: %v", err.Error())
}
}
for _, k := range s.config.MetaAsTags {
s.mp.AddMoveMetaToTags("true", k, k)
}
precision := influx.Second
if len(s.config.Precision) > 0 {
switch s.config.Precision {
case "s":
precision = influx.Second
case "ms":
precision = influx.Millisecond
case "us":
precision = influx.Microsecond
case "ns":
precision = influx.Nanosecond
}
}
// Create http client
s.client = &http.Client{
Transport: &http.Transport{
MaxIdleConns: 1, // We will only ever talk to one host.
IdleConnTimeout: s.config.idleConnTimeout,
},
Timeout: s.config.timeout,
}
// Configure influx line protocol encoder
s.encoder.SetPrecision(precision)
return s, nil
}

44
sinks/httpSink.md Normal file
View File

@@ -0,0 +1,44 @@
## `http` sink
The `http` sink uses POST requests to a HTTP server to submit the metrics in the InfluxDB line-protocol format. It uses JSON web tokens for authentification. The sink creates batches of metrics before sending, to reduce the HTTP traffic.
### Configuration structure
```json
{
"<name>": {
"type": "http",
"url" : "https://my-monitoring.example.com:1234/api/write",
"jwt" : "blabla.blabla.blabla",
"username": "myUser",
"password": "myPW",
"timeout": "5s",
"idle_connection_timeout" : "5s",
"flush_delay": "2s",
"batch_size": 1000,
"precision": "s",
"process_messages" : {
"see" : "docs of message processor for valid fields"
},
"meta_as_tags" : []
}
}
```
- `type`: makes the sink an `http` sink
- `url`: The full URL of the endpoint
- `jwt`: JSON web tokens for authentication (Using the *Bearer* scheme)
- `username`: username for basic authentication
- `password`: password for basic authentication
- `timeout`: General timeout for the HTTP client (default '5s')
- `max_retries`: Maximum number of retries to connect to the http server
- `idle_connection_timeout`: Timeout for idle connections (default '120s'). Should be larger than the measurement interval to keep the connection open
- `flush_delay`: Batch all writes arriving in during this duration (default '1s', batching can be disabled by setting it to 0)
- `batch_size`: Maximal batch size. If `batch_size` is reached before the end of `flush_delay`, the metrics are sent without further delay
- `precision`: Precision of the timestamp. Valid values are 's', 'ms', 'us' and 'ns'. (default is 's')
- `process_messages`: Process messages with given rules before progressing or dropping, see [here](../pkg/messageProcessor/README.md) (optional)
- `meta_as_tags`: print all meta information as tags in the output (deprecated, optional)
### Using `http` sink for communication with cc-metric-store
The cc-metric-store only accepts metrics with a timestamp precision in seconds, so it is required to use `"precision": "s"`.

Some files were not shown because too many files have changed in this diff Show More