Kubernetes pod stuck in CrashLoopBackOff (container restarts with increasing delay). How do I find out why it's crashing and fix it?

Cause: CrashLoopBackOff isn't the error itself. It means a container in the pod keeps exiting and the kubelet is restarting it with an increasing back-off delay (up to five minutes). The real reason is in the container's exit code and logs. Common causes: the app crashes on startup (missing environment variable, config file, secret or database connection), a wrong command/args or entrypoint, the process exits immediately because it isn't a long-running server, OOMKilled because memory limits are too low, a liveness probe that fails and kills a healthy-but-slow app, or missing permissions (read-only filesystem, non-root user). How to find the reason (Kubernetes' debugging guide): 1. kubectl describe pod <pod>: look at Last State (Reason, Exit Code), and Events at the bottom (probe failures, OOMKilled, image issues). 2. kubectl logs <pod> --previous (add -c <container> for multi-container pods): shows the output of the crashed run, not the current restart. 3. Exit code hints: 1 = app error (read the logs); 137 = killed (OOMKilled or a failed liveness probe); 139 = segfault; 127/126 = command not found / not executable (check command and image). Fix (by cause): - App config errors: check the ConfigMap/Secret names and keys referenced in env/envFrom and volume mounts exist in the same namespace. - OOMKilled: raise resources.limits.memory or reduce the app's memory use; set requests realistically. - Liveness probe kills: add a startupProbe or a longer initialDelaySeconds/failureThreshold so slow starts aren't treated as failures; make sure the probe path and port are right. - Wrong command: compare command/args with the image's ENTRYPOINT/CMD; test the image locally with docker run. - To inspect the container interactively, use kubectl debug (an ephemeral debug container or a copy of the pod with a different command) instead of editing the production image. Verify: kubectl get pod shows Running with the RESTARTS count no longer increasing, and kubectl describe shows no new failure events. Sources: https://kubernetes.io/docs/tasks/debug/debug-application/debug-pods/ https://kubernetes.io/docs/concepts/workloads/pods/pod-lifecycle/