CrashLoopBackOff: read the exit code before you touch the YAML

verbatim errorBack-off restarting failed container pod/app-7d9c6b5f4-xqz2l: CrashLoopBackOff container "app" in pod "app-7d9c6b5f4-xqz2l" is waiting to start: CrashLoopBackOff

Problem

Back-off restarting failed container
pod/app-7d9c6b5f4-xqz2l: CrashLoopBackOff

The reflex is to re-read the deployment YAML for ten minutes, then kubectl delete pod and hope. The pod came back, crashed again, back-off stretched to 5 minutes, and the on-call lost an hour.

Root cause

CrashLoopBackOff is not an error, it is a state: the container starts and exits non-zero, repeatedly. Six causes cover nearly every case, and the exit code tells you which:

1: the app itself failed — bad config, missing env var, unhandled exception
137 (SIGKILL): OOMKilled or a liveness probe kill; check RESTARTS and kubectl describe
126/127: command not executable / not found — image vs command mismatch
0 with restarts: the process ran and exited cleanly — usually a foreground flag missing (a server started in background mode)

There is a diagnostic order that finds the cause in under two minutes.

fix preview — first 3 of 6 lines (bash), truncated:

kubectl describe pod app-7d9c6b5f4-xqz2l | grep -A8 "Last State"
# Last State:       Terminated
#   Reason:       Error
… 3 more lines in the fix

🔒 the fix — including 3 code blocks — is members-only. $1/mo unlocks everything.