Golang:实现http服务器健康检查。 gocraft /健康

时间:2015-06-02 11:47:50

标签: go http-headers metrics endpoints endpointbehavior

我想检查我的服务的健康状况,了解每个endPoint的指标。 我的服务调用一些其他服务并收到一个Json代码,我用它制作模板,然后我把它发送到http.ResponseWriter。

我搜索过,我发现这个包“gocraft / health”,但我真的不明白它是如何工作的。

是否有任何其他方式或包来生成指标,或者我应该只使用“gocraft / health。

提前谢谢

2 个答案:

答案 0 :(得分:2)

最后,我选择" gocraft / health" ,这是一个很棒的图书馆。

使用示例:

package main

import (
    "log"
    "net/http"
    "os"
    "time"

    "github.com/gocraft/health"
)

//should be global Var
var stream = health.NewStream()

func main() {
    // Log to stdout!
    stream.AddSink(&health.WriterSink{os.Stdout})
    // Make sink and add it to stream
    sink := health.NewJsonPollingSink(time.Minute*5, time.Minute*20)
    stream.AddSink(sink)
    // Start the HTTP server! This will expose metrics via a JSON API.
    adr := "127.0.0.1:5001"
    sink.StartServer(adr)

    http.HandleFunc("/api/getVastPlayer", vastPlayer)
    log.Println("Listening...")
    panic(http.ListenAndServe(":2001", nil))
}

根据上面的初始化选项,您的指标会以5分钟的方式聚合。我们会在内存中保留20分钟的数据。什么都没有持久到磁盘。

您可以根据需要创建任意数量的作业

func vastPlayer(w http.ResponseWriter, r *http.Request) {
  job_1 := stream.NewJob("/api/getVastPlayer")

  ...
  ...

  if bol {
    job_1.Complete(health.Success)
  } else {
    job_1.Complete(health.Error)
  }
}

启动应用后,这将通过JSON API公开指标。您可以浏览/health端点(例如127.0.0.1:5001/health)以查看指标。你会得到类似的东西:

{
  "instance_id": "sd-69536.29342",
  "interval_duration": 86400000000000,
  "aggregations": [
    {
      "interval_start": "2015-06-11T02:00:00+02:00",
      "serial_number": 1340,
      "jobs": {
        "/api/getVastPlayer": {
          "timers": {},
          "events": {},
          "event_errs": {},
          "count": 1328,
          "nanos_sum": 140160794784,
          "nanos_sum_squares": 9.033775178022173E+19,
          "nanos_min": 34507863,
          "nanos_max": 2736850494,
          "count_success": 62,
          "count_validation_error": 1266,
          "count_panic": 0,
          "count_error": 0,
          "count_junk": 0
        },
        "timers": {},
        "events": {},
        "event_errs": {}
      }
    }
  ]
}

有关更多信息和功能,请查看以下链接:

https://github.com/gocraft/health

答案 1 :(得分:0)

如果您由于要公开/health端点而遇到此问题,那么即将进行健康检查的RFC:https://github.com/inadarei/rfc-healthcheck

还有一个Go库health-go,用于公开符合该RFC的健康端点:https://github.com/nelkinda/health-go

示例:

package main

import (
    "github.com/nelkinda/health-go"
    "net/http"
)

func main() {
    // 1. Create the health Handler.
    h := health.New(health.Health{Version: "1", ReleaseID: "1.0.0-SNAPSHOT"}) 

    // 2. Add the handler to your mux/server.
    http.HandleFunc("/health", h.Handler)

    // 3. Start your server.
    http.ListenAndServe(":80", nil)
}

它是可扩展的,并支持许多内置检查,例如正常运行时间和sysinfo。

免责声明:我是health-go的作者。