go-runtime-metrics

Collect golang runtime metrics, pushing to InfluxDB or pulling with Telegraf. Inspired by https://github.com/bmhatfield/go-runtime-metrics

Installation

go get -u github.com/tevjef/go-runtime-metrics

Push Usage

This library can be configured to push metrics directly to InfluxDB.

import (
	metrics "github.com/tevjef/go-runtime-metrics"
)

func main() {
	err := metrics.RunCollector(metrics.DefaultConfig)
	
	if err != nil {
	   // handle error
	}
}
	

Once imported and running, you can expect a number of Go runtime metrics to be sent to InfluxDB. An example of what this looks like when configured to work with Grafana:

Download Dashboard

Pull Usage via expvar

Package expvar provides a standardized interface to public variables. This library provides an exported InfluxDB formatted variable with a few other benefits:

• Metric names are easily parsed by regexp.
• Lighter than the standard library memstat expvar
• Includes stats for cpu.cgo_calls, cpu.goroutines and timing of the last GC pause with mem.gc.pause.
• Works out the box with Telegraf's InfluxDB input plugin

Import this library's expvar package with import _ "github.com/tevjef/go-runtime-metrics/expvar" to export a variable with default configurations.

{
  "/go/bin/binary": {
    "name": "go_runtime_metrics",
    "tags": {
      "go.arch": "amd64",
      "go.os": "darwin",
      "go.version": "go1.7.4"
    },
    "values": {
      "cpu.count": 4,
      "cpu.cgo_calls": 1,
      "cpu.goroutines": 2,
      "mem.alloc": 667576,
      "mem.frees": 104,
      "mem.gc.count": 0,
      "mem.gc.last": 0,
      "mem.gc.next": 4194304,
      "mem.gc.pause": 0,
      "mem.gc.pause_total": 0,
      "mem.gc.sys": 65536,
      "mem.heap.alloc": 667576,
      "mem.heap.idle": 475136,
      "mem.heap.inuse": 1327104,
      "mem.heap.objects": 5227,
      "mem.heap.released": 0,
      "mem.heap.sys": 1802240,
      "mem.lookups": 3,
      "mem.malloc": 5331,
      "mem.othersys": 820558,
      "mem.stack.inuse": 294912,
      "mem.stack.mcache_inuse": 4800,
      "mem.stack.mcache_sys": 16384,
      "mem.stack.mspan_inuse": 14160,
      "mem.stack.mspan_sys": 16384,
      "mem.stack.sys": 294912,
      "mem.sys": 3018752,
      "mem.total": 667576
    }
  }
}

Configuring with Telegraf

Your program must import _ "github.com/tevjef/go-runtime-metrics/expvar in order for an InfluxDB formatted variable to be exported via /debug/vars.

1. Install Telegraf

2. Make a config file utilizing the influxdb input plugin and an output plugin of your choice.

   [[inputs.influxdb]]
     urls = ["http://localhost:6060/debug/vars"]
   
   [[outputs.influxdb]]
     urls = ["http://localhost:8086"]
     ## The target database for metrics (telegraf will create it if not exists).
     database = "stats" # required
     
   ## [[outputs.file]]
   ##   files = ["stdout"]
   ##   data_format = "json"
   

3. Start the Telegraf agent with telegraf -config config.conf

Benchmarks

Benchmark against standard library memstat expvar:

$ go test -bench=. -parallel 16 -cpu 1,2,4

BenchmarkMetrics          100000             12456 ns/op            4226 B/op         21 allocs/op
BenchmarkMetrics-2         20000             63597 ns/op            4264 B/op         21 allocs/op
BenchmarkMetrics-4         50000             28797 ns/op            4266 B/op         21 allocs/op
BenchmarkMemstat           20000             78009 ns/op           52264 B/op         12 allocs/op
BenchmarkMemstat-2         10000            155930 ns/op           52264 B/op         12 allocs/op
BenchmarkMemstat-4         10000            144849 ns/op           52266 B/op         12 allocs/op

  System Info: 

  Processor Name:	Intel Core i5
  Processor Speed:	3.5 GHz
  Number of Processors:	1
  Total Number of Cores:	4
  L2 Cache (per Core):	256 KB
  L3 Cache:	6 MB
  Memory:	32 GB
  Bus Speed:	400 MHz