What to do if Kubernetes crashes?

If Kubernetes has crashed, you should:

1. Check the status of the main cluster components:

   • kubectl get nodes — check node availability.
   • kubectl get pods -A — check the status of system pods (kube-system).

2. Check component logs:

   • Logs of kube-apiserver, kube-controller-manager, kube-scheduler on master nodes.
   • Logs of kubelet on worker nodes.

3. Check the status of etcd (if using a dedicated etcd cluster), as it stores the cluster state.

4. Check network settings and CNI plugins, as network issues often cause service failures.

5. If the cluster is deployed via a managed service (e.g., GKE, EKS), check the status in the management panel.

6. Restart failed components or nodes if necessary.

7. Use monitoring tools (Prometheus, Grafana) for diagnostics.

It is important to have a recovery plan and etcd backups for critical cases.