Kubernetes troubleshooting guides

The incidents the Kubernetes Simulator's missions are built around, written down: what each error means, how to find the cause with kubectl, and how to fix it. Every output shown was captured from a real Kubernetes v1.35 cluster.

• How to fix CrashLoopBackOff in KubernetesCrashLoopBackOff means a container keeps exiting and Kubernetes waits longer between restarts. Read the logs, check the exit code, then fix the cause.
• ImagePullBackOff and ErrImagePull: causes and fixesErrImagePull and ImagePullBackOff mean the kubelet cannot pull a pod's image. Find the reason in the pod's events, fix the image or credentials, or roll back.
• OOMKilled (exit code 137): fixing out-of-memory podsOOMKilled with exit code 137 means a container went over its memory limit and was killed. Confirm it with kubectl describe, measure usage, then set the limit.
• Roll back a bad deployment with kubectl rollout undoUse kubectl rollout undo to put a Deployment back on its last working revision, or a chosen one with --to-revision, and check the rollback has finished.
• Debugging NetworkPolicy: why pods can't reach each otherWhen a NetworkPolicy blocks traffic, connections time out silently. Find the policies that select the pod, check ports and DNS, and test with curl timeouts.
• Practising for the CKA and CKAD with a Kubernetes simulatorHow the Kubernetes Simulator's missions line up with the CKA and CKAD exam domains, what they cover well, and what you still need a real cluster for.

Reading is half of it. Fix them yourself in the Kubernetes Simulator, free in your browser.