Deployment Log Persistence for Immediate Startup Failures

Answer to the Question

Replit does not expose a separate system‑level log stream that records events occurring before the runtime’s logging agent attaches. The Deployment Logs tab only shows stdout and stderr after the agent is connected. Therefore, if a container exits immediately on start‑up, the logs tab may stay empty.

Likely Explanation

When a deployment starts, Replit launches a lightweight container and then attaches a logging agent to the process’s streams. If the main process terminates during the very first boot cycle—before the agent can hook in—the output never reaches the UI. Replit’s platform does not provide a kernel‑level or init‑log view for these early failures, so the only observable indicator is the container’s exit code.

Recommended Verification Steps

1. Add an early print statement at the top of your entrypoint (e.g., print('startup') in Python or echo "startup" in a shell script). Redeploy and check if the first line appears in the logs tab; if it does, the crash happens after the agent attaches.
2. Use the Replit CLI for verbose output:

   replit deploy --verbose
   replit logs --tail
   

   This shows the last few seconds of console output and may surface hidden error messages.
3. Inspect the container’s exit code via the UI or CLI. A non‑zero exit code (commonly 1–125) confirms a crash. If the logs tab remains empty, the exit code itself is the primary clue.
4. Reproduce locally by running replit run or launching a Docker image that matches the deployment environment. Capture the exit code and any console output; this often mirrors the production failure.
5. Set a debug flag (e.g., REPLIT_DEBUG=1) or enable a minimal health‑check script that writes to stdout before the main process starts. This forces the logging agent to attach early.

Additional Tips

• Verify that your entrypoint is correctly configured; an accidental exec of a background process can terminate the main process before logging begins.
• Check resource limits (CPU, memory) in the deployment settings; a container can exit immediately if it exceeds quotas.
• If the logs tab remains empty after these steps, consider contacting Replit support with the repository URL and the observed exit code.