Worker pods exit immediately due to incorrect container command

Worker pods show Completed status instead of Running. Pods start, execute a one-time script (like database migrations), and then exit. The container command runs a short-lived process instead of the long-running Celery worker process.

Inspect deployment container command with kubectl get deployment. Ensure command is set to run Celery worker as a long-running process: command: ["celery"] with args including -A, worker, --loglevel, and --concurrency flags.