Introduction to Kubernetes Operators

  • What: Operators are software extensions that manage operational complexity of applications on Kubernetes.
  • Components: Custom Resource (CR), Custom Resource Definition (CRD), and Controller.
  • Why: To automate application-specific tasks like backup, updates, and scaling.

Writing Custom Operators

  • Decide Use-Case: Determine if your application has unique complexities that Kubernetes cannot handle natively.
  • Design CRDs: Outline the Custom Resource Definitions that capture the desired state of your application.
  • Implement Controller: Write the controller logic to watch the CR and reconcile the actual state to match the desired state.
  • Languages: Go is the most commonly used, but you can use Python, Java, etc.

Example: Custom Operator for a Redis Cache in Java

Here’s a very simplified example written in Java using the Fabric8 Kubernetes Client.

import io.fabric8.kubernetes.client.KubernetesClient;
import io.fabric8.kubernetes.client.Watcher;
import io.fabric8.kubernetes.client.Watcher.Action;

public class RedisOperator {

    private KubernetesClient client;

    public RedisOperator(KubernetesClient client) {
        this.client = client;

    public void watchRedisResources() {
            .watch(new Watcher<RedisCustomResource>() {
                public void eventReceived(Action action, RedisCustomResource resource) {
                public void onClose(KubernetesClientException e) {
                    // Handle close events

    public void reconcile(RedisCustomResource redisResource) {
        String name = redisResource.getMetadata().getName();
        String namespace = redisResource.getMetadata().getNamespace();
        // Reconciliation logic here, for example, create a new Pod

Operator Framework

  • Operator SDK: Provides tools to generate operator boilerplate code, build and push container images.
  • Kubebuilder: Another toolkit that simplifies creating controllers and CRDs.
  • Metacontroller: Allows you to create simple operators using scripts in languages like JavaScript and Python.

Deploying and Managing Operators

  1. Install Operator: Apply the CRD and deploy the operator using a Deployment manifest.
  2. Deploy CR: Create a Custom Resource that matches the CRD to trigger the operator’s controller.
  3. Manage: Update the CR to change application configuration, scale, etc.
  4. Monitor: Use Kubernetes native tools (kubectl logs, kubectl describe) to debug issues.