Unexpected container hibernation during Replit deployment workflow

Answer to the core question

The production instance created by Replit’s deployment workflow does inherit the same free‑tier idle timeout as the development container. After roughly 30 minutes of no inbound traffic, the container will hibernate (auto‑stop), making the deployed app unreachable until the next request triggers a cold start.

Confirmed facts

• Free‑tier Replit containers hibernate after ~30 minutes of user inactivity (research).
• There is no official setting or feature flag for free‑tier users to disable or extend this timeout.
• The only built‑in way to avoid hibernation is to upgrade to a paid plan that provides "Always On" instances.

Likely explanation

Replit treats the deployed container as any other container on the free tier: it monitors network activity. If the app does not receive any HTTP (or TCP) requests, the platform marks it idle and eventually stops it to conserve resources. The .replit run command and Secrets configuration do not influence this monitoring.

Steps to mitigate hibernation on the free tier

1. Ensure your app exposes an HTTP endpoint that can be pinged (e.g., a simple /ping route returning 200 OK).
2. Add a lightweight keep‑alive mechanism that calls this endpoint at intervals shorter than the timeout (e.g., every 20 minutes). This can be done:
   • Inside the app: setInterval(() => fetch('/ping'), 20 * 60 * 1000);
   • Externally: use a free cron service (e.g., cron-job.org) or another Repl that pings your deployed URL.
3. Verify that the keep‑alive requests are reaching the container by checking your app’s logs or adding a timestamp to the /ping response.
4. If you prefer not to modify the app, consider upgrading to a paid Replit plan, which provides an "Always On" toggle in the deployment settings that disables the auto‑stop behavior.

Missing diagnostic detail

To confirm whether the above steps will work for your specific case, please let me know: Does your deployed app already expose an HTTP endpoint that can be reached from outside the container? If the answer is no, you will need to add one (as shown in step 1) before the keep‑alive ping can be effective.