Kubernetes Pod Startup & CrashLoop Errors
The complete set of "my pod won't start" failures: crash loops, image-pull failures, OOM kills, bad container config, failing init containers, and node evictions. Each guide leads with the one-line cause and the `kubectl logs --previous` / `describe` commands that prove it. Jump to your symptom below, or hand the logs to the assistant.
Fix your error now
Paste the error or logs and let the Incident Assistant diagnose it, or run a guided triage prompt with AI — no copy-paste.
Every guide in this cluster
11 guides-
'CrashLoopBackOff' Pod Restart Loop
Fix the CrashLoopBackOff pod restart loop in Kubernetes: diagnose startup crashes, failing liveness probes, missing dependencies, OOM during init, and bad entrypoints.
-
'ImagePullBackOff' / 'ErrImagePull' Image Pull Failures
Fix ImagePullBackOff and ErrImagePull in Kubernetes: diagnose bad image names, missing tags, private registry secrets, expired credentials, rate limits, and DNS.
-
'OOMKilled' Exit Code 137 Out-of-Memory Kills
Fix OOMKilled (exit code 137) in Kubernetes: diagnose low memory limits, leaks, JVM/Node heaps ignoring cgroups, batch spikes, greedy sidecars, and node pressure.
-
'CreateContainerConfigError' Missing ConfigMap or Secret
Fix the Kubernetes CreateContainerConfigError: diagnose missing ConfigMaps and Secrets, absent keys, wrong namespaces, ordering races, and bad envFrom/valueFrom references.
-
'CreateContainerError' Runtime Create Failure
Fix the CreateContainerError: resolve bad commands, missing host mounts, device conflicts, and runtime issues that stop the container runtime from creating the container.
-
'ErrImageNeverPull' Missing Local Image
Fix the ErrImageNeverPull error: load the image onto the node, correct imagePullPolicy Never, and align tags so pods using preloaded images start cleanly.
-
'ErrImagePull' First-Attempt Image Pull Failure
Fix ErrImagePull in Kubernetes: diagnose wrong image names, nonexistent tags, unreachable registries, missing imagePullSecrets, and expired credentials fast.
-
'Evicted: The node was low on resource' Pod Eviction
Fix Evicted 'The node was low on resource: ephemeral-storage/memory': stop disk and memory pressure, set requests and limits, and prevent kubelet node-pressure evictions.
-
'Init Container Failed' — Unblock Init:Error
Fix an init container that exits non-zero in Kubernetes: read Init:Error and Init:CrashLoopBackOff, decode exit codes, and check init logs, dependencies, and mounts.
-
'Readiness probe failed: context deadline exceeded' Probe Timeouts
Fix Kubernetes probe timeouts: 'Readiness probe failed: context deadline exceeded' and liveness restart loops via initialDelaySeconds, timeoutSeconds, and paths.
-
'Readiness probe failed: statuscode: 503' — Fix Unready Pods
Fix the Kubernetes 'Readiness probe failed: statuscode: 503' error: diagnose dependency checks, warmup timing, and probe endpoints that report unready.
Kubernetes & Helm AI prompts
Copy-paste, production-safe prompts for this stack.
Kubernetes & Helm command center
Top errors, validators, and runbooks for the whole stack.
Frequently asked questions
How do I find why a pod is in CrashLoopBackOff?
kubectl logs <pod> --previous to read the last run before the restart, and kubectl describe pod <pod> for the termination reason and exit code. Exit 137 is an OOM kill, exit 1/2 is usually an app or config error. See CrashLoopBackOff.What is the difference between ImagePullBackOff and ErrImagePull?
ErrImagePull is the first failed pull attempt; ImagePullBackOff is Kubernetes backing off and retrying it. Both mean the kubelet can't fetch the image — a wrong name/tag, a missing imagePullSecret, expired registry credentials, or rate limiting. See ImagePullBackOff.