Google 云监控 - 警报政策：状态代码：409 - 对项目配置的并发编辑过多

Question

我正在尝试为Google Cloud Platform中的Kubernetes Clusters创建警报策略。下面是示例代码。service, err := monitoring.NewService(context.Background())if err != nil {    log.Panicln(err)    return}mqlCondition := &monitoring.MonitoringQueryLanguageCondition{    Duration: "60s",    Query: `fetch k8s_pod            | metric 'kubernetes.io/pod/volume/utilization'            | filter                (resource.cluster_name == 'test'                 && resource.pod_name =~ 'server.*')                && (metric.volume_name =~ 'dat.*')            | align mean_aligner()            | window 10m            | condition value.utilization > 0.001 '10^2.%'            `,    Trigger: &monitoring.Trigger{        Count: 1,    },}condition := monitoring.Condition{    DisplayName:                      "MQL-based Condition",    ConditionMonitoringQueryLanguage: mqlCondition,}alertpolicy := &monitoring.AlertPolicy{    DisplayName:          "Prakash1",    Combiner:             "OR",    Conditions:           []*monitoring.Condition{&condition},    NotificationChannels: []string{"projects/abc-app/notificationChannels/16000000099515524778"},    }p, err := service.Projects.AlertPolicies.Create("projects/abc-app", alertpolicy).Context(context.Background()).Do()if err != nil {    log.Panicln(err)    return}当我同时创建两个或多个警报策略时，我收到以下错误："googleapi: Error 409: Too many concurrent edits to the project configuration. Please try again., aborted"你能告诉我如何解决这个错误吗？

Answer 1

如错误所述，您无法同时添加策略。我建议您并行生成创建请求并序列化实际的 API 调用。您可以通过缓冲通道和多个 go 例程来实现此目的。例如：

var(

  backlogSize = 3 //change as per your needs

  requests = make(chan *monitoring.AlertPolicy, backlogSize)

)

func createPolicies(){

  ...

  go func(){

     //init the service

     ...

     for policy := range requests {

       p, err := service.Projects.AlertPolicies.Create("projects/abc-app", policy).Context(context.Background()).Do()

       if err != nil {

         log.Println(err)

       }

     }

  }()

  go func(){

     newPolicy := &monitoring.AlertPolicy

     //fill policy

     ...

     requests <- newPolicy 

  }()

  ...

  //wait for completion and close requests channel

}

另一种解决方案是使用指数退避+抖动重试每个失败的并发请求