Heapster-client

Introduction

Heapster enables Container Cluster Monitoring and Performance Analysis. And it has some APIs. This project is a library that makes it easier for you to use the APIs.

NB: If you are not familiar to the Heapster model, I suggest you read Heapster Metric Model first.

Usage

1. Get the library

   go get github.com/niyanchun/heapster-client
   

2. List Cluster-Level metric, and get cpu/usage_rate metric statistics:

   package main
   
   import (
       "fmt"
       "github.com/niyanchun/heapster-client/client"
       "time"
       "github.com/niyanchun/heapster-client/type/v1"
   )
   
   func main() {
       HEAPSTER_URL := "http://192.168.56.101:8080/api/v1/proxy/namespaces/kube-system/services/heapster"
       client := client.NewClient(HEAPSTER_URL, "", "")
   
       start, end := getTimeRange()
       list, err := client.ListClusterMetric()
       checkErr(err)
       fmt.Printf("ListClusterMetric:\n %v", list)
   
       metrics, err := client.GetClusterMetrics(v1.CPU_USAGE_RATE, start, end)
       checkErr(err)
       fmt.Printf("ListClusterMetric:\n %v", metrics)
   }
   
   func checkErr(err error) {
       if err != nil {
           panic(err)
       }
   }
   
   func getTimeRange() (string, string) {
       now := time.Now().UTC()
       end := now.Format(time.RFC3339)
       start := now.Add(-600 * 1e9).Format(time.RFC3339) // ten minutes before
       fmt.Println(start, end)
   
       return start, end
   }
   
   

Compatibility

This library is test under Kubernetes 1.6.0 and Heapster v1.3.0-beta.0.