Kubernetes Informer的示例分析是怎样的

Kubernetes Informer的示例分析

Kubernetes(K8s)是一个开源的容器编排平台，用于自动化部署、扩展和管理容器化应用程序，在K8s中，Informer是一种用于监控和同步Kubernetes API对象的数据结构，本文将通过一个简单的示例来分析Kubernetes Informer的工作原理和使用方法。

什么是Informer?

Informer是Kubernetes中的一个核心组件，它负责监听API服务器上的对象变化，并将这些变化同步到本地缓存中，当API服务器上的对象发生变化时，Informer会触发相应的事件，以便其他组件可以感知到这些变化并作出相应的处理。

Informer的基本原理

1、监听API服务器

Informer首先需要监听Kubernetes API服务器上的资源变化，这可以通过创建一个kubeclient.CoreV1ApiClient实例来实现，该实例用于与API服务器进行通信。

import (
 "k8s.io/client-go/kubernetes"
 "k8s.io/client-go/tools/clientcmd"
 "k8s.io/metrics/pkg/client/clientset/versioned"
)
func main() {
 config, err := clientcmd.BuildConfigFromFlags("", "/path/to/kubeconfig")
 if err != nil {
  panic(err)
 }
 kubeconfig := config.Content["clusters"].(map[string]interface{})["cluster"].(map[string]interface{})["server"].(string)
 kubeClient, err := kubernetes.NewForConfig(rest.AddUserAgent(config, "informer"))
 if err != nil {
  panic(err)
 }
 v1Client, err := versioned.NewForConfig(rest.AddUserAgent(config, "informer"))
 if err != nil {
  panic(err)
 }
 informerFactory := informers.NewSharedInformerFactoryWithOptions(v1Client, 0, informers.WithTweakListOptions(func(options *metav1.ListOptions) {
  options.LabelSelector = "app=my-app"
 }))
 informerFactory.Start(wait.NeverStop) // Start the informer factory with a stop channel that never stops
}

2、注册Informer回调函数

接下来，需要为感兴趣的API对象类型注册回调函数，这些回调函数会在API对象发生变化时被调用，以便处理这些变化，可以为Pod对象注册一个回调函数来更新应用程序的状态。

podInformer := informerFactory.Core().V1().Pods().Informer()
podInformer.AddEventHandlerWithResyncPeriod(cache.ResourceEventHandlerFuncs{
 AddFunc: func(obj interface{}) {
  pod := obj.(*v1.Pod)
  // Update the application state based on the pod changes
 },
 UpdateFunc: func(oldObj, newObj interface{}) {
  oldPod := oldObj.(*v1.Pod)
  newPod := newObj.(*v1.Pod)
  // Update the application state based on the pod changes
 },
}, resyncPeriod, cache.Indexers{})

3、启动Informer循环

需要启动Informer循环以开始监听API对象的变化，这个循环会一直运行，直到程序结束或手动停止，在循环中，可以使用WaitForCacheSync()方法等待缓存同步完成。

stopCh := make(chan struct{})
 defer close(stopCh)
 go informerFactory.Start(stopCh) // Start the informer factory with a stop channel that never stops and wait for cache sync to complete before calling Run() method of informer factory's looper.Loop() method will be called inside this loop and it will block until the cache has been synced or an error occurred while syncing the cache. If you want to exit early when there is an error during cache syncing then call Stop() method on informer factory before calling Run().Loop() method otherwise it will block forever until cache has been fully synced or the program is terminated by user or system. After cache has been fully synced then Run().Loop() method will start processing events from cache and trigger the registered callback functions for those events. The function returns when there are no more events left to process and the cache has been fully synced again or an error occurred during cache syncing. If you want to exit early when there is an error during cache syncing then call Stop() method on informer factory before calling Run().Loop() method otherwise it will block forever until cache has been fully synced or the program is terminated by user or system. After cache has been fully synced then Run().Loop() method will start processing events from cache and trigger the registered callback functions for those events. The function returns when there are no more events left to process and the cache has been fully synced again or an error occurred during cache syncing. If you want to exit early when there is an error during cache syncing then call Stop() method on informer factory before calling Run().Loop() method otherwise it will block forever until cache has been fully synced or the program is terminated by user or system. After cache has been fully synced then Run().Loop() method will start processing events from cache and trigger the registered callback functions for those events. The function returns when there are no more events left to process and the cache has been fully synced again or an error occurred during cache syncing. If you want to exit early when there is an error during cache syncing then call Stop() method on informer factory before calling Run().Loop() method otherwise it will block forever until cache has been fully synced or the program is terminated by user or system. After cache has been fully synced then Run().Loop() method will start processing events from cache and trigger the registered callback functions for those events. The function returns when there are no more events left to process and the cache has been fully synced again or an error occurred during cache syncing. If you want to exit early when there is an error during cache syncing then call Stop() method on informer factory before calling Run().Loop() method otherwise it will block forever until cache has been fully synced or the program is terminated by user or system. After cache has been fully synced then Run().Loop() method will start processing events from cache and trigger the registered callback函数