goappmonitor
Golang application performance data monitoring.
GoAppMonitor is a library which provides a monitor on your golang applications. It contains system level based monitoring and business level monitoring(custom monitoring).Just add the repository into your apps and register what you want to monitoring.
Summary
Using GoAppMonitor to monitor the golang applications, in general as following:
In your golang application code, the user calls the statistics function provided by goappmonitor; when the statistics function is called, the appmonitor generates a statistical record, and is stored in memory.GoAppMonitor will automatically and regularly record these statistics push to the agent such as Open-Falcon agent.
Version
Current version support:
-
v0.0.2
- Open-Falcon (Open source monitoring system of Xiaomi)
- InfluxDB (Scalable datastore for metrics, events, and real-time analytics)
todo....
- support more agent frameworks,such as elasticsearch...
- go processes manager and debug online...
Install
go get github.com/wgliang/goappmonitor
Demo
Usage
Below is an example which shows some common use cases for goappmonitor. Check example for more usage.
Detail API
package main
import (
"math/rand"
"time"
appm "github.com/wgliang/goappmonitor"
)
// Base or system performance data,such as memeory,gc,network and so on.
func baseOrsystem() {
for _ = range time.Tick(time.Second * time.Duration(10)) {
// (commonly used) Meter, used to sum and calculate the rate of change. Use scenarios
// such as the number of home visits statistics, CG etc..
pv := int64(rand.Int31n(100))
appm.Meter("appm.meter", pv)
appm.Meter("appm.meter.2", pv-50)
// (commonly used) Gauge, used to preserve the value of the instantaneous value of the
// type of record. Use scenarios such as statistical queue length, statistics CPU usage,
// and so on.
queueSize := int64(rand.Int31n(100) - 50)
appm.Gauge("appm.gauge", queueSize)
cpuUtil := float64(rand.Int31n(10000)) / float64(100)
appm.GaugeFloat64("appm.gauge.float64", cpuUtil)
}
}
// Custom or business performance data,such as qps,num of function be called, task queue and so on.
func customOrbusiness() {
for _ = range time.Tick(time.Second) {
// Histogram, using the exponential decay sampling method, the probability distribution of
// the statistical object is calculated. Using scenarios such as the probability distribution
// of the statistics home page to access the delay
delay := int64(rand.Int31n(100))
appm.Histogram("appm.histogram", delay)
}
}
func main() {
var ch chan int
go baseOrsystem()
go customOrbusiness()
<-ch
}
Credits
Repository is base on goperfcounter of niean
Logo is desigend by xuri