Docker container keeps restarting — how to debug

If a Docker container keeps restarting, the process inside is crashing—not “Docker being flaky.” Read logs and exit codes before changing restart policies or rebuilding images.

Debug steps

1. Logs

   docker ps -a
   docker logs CONTAINER_NAME --tail 100

2. Exit code

   docker inspect CONTAINER_NAME --format '{{.State.Status}} exit={{.State.ExitCode}} error={{.State.Error}}'

   Exit 137 often means OOM killed. 1 is generic app failure.

3. Run interactively (bypass restart loop)

   docker run -it --rm --entrypoint /bin/sh IMAGE

   Or override command in compose temporarily to sleep infinity and exec in.

4. Check compose / run config

   • Missing DATABASE_URL, wrong volume mount path
   • Port already in use on host
   • Healthcheck too aggressive before app is ready

5. Resources

   docker stats --no-stream
   dmesg | tail   # OOM killer messages

Common causes

Clue	Fix
”connection refused” to DB	Start DB first, fix network alias
”permission denied” on volume	UID/GID mismatch, fix ownership
Works locally, fails on VPS	Wrong arch image, missing .env on server
OOM 137	Increase memory limit or fix leak

Production caution

Do not docker system prune -a on a live box to “fix” things. Do not remove volumes without knowing their names.

Get help without YOLO commands

Ohuriya AI is an AI DevOps Copilot—you can paste docker logs output and approve diagnostic commands one at a time. Connect your VPS · disk full fixes