Container on App Service keeps getting stopped and terminated

When a custom container on Linux App Service terminates only a few seconds after start and the log shows:

Container is running.

Container start method finished…

Container is terminating. Grace period: 0 seconds.

it usually means the entrypoint process exited, even if the platform briefly reports “running.”

Key causes to check:

1. Your container’s main process is stopping immediately
   App Service requires the entrypoint to stay alive. If your .NET app runs, prints logs, and then exits (for example, missing a web server binding), the container is deleted instantly.
2. The app is not listening on the required port
   Linux App Service expects the app to listen on port 8080 (or the port defined in WEBSITES_PORT).
   If the container listens on 5000/80/etc. without exposing 8080, the health probe fails and App Service stops the container.
3. No “long-running” foreground process
   If the app launches and then returns control to the shell (for example, backgrounding the process), App Service treats it like a completed job and terminates it.
4. Dockerfile CMD/ENTRYPOINT mismatch
   A misconfigured Dockerfile can cause the runtime to exit instantly:

• ENTRYPOINT overwritten by App Service
• Using CMD ["dotnet", "myapp.dll"] but the DLL is missing or misnamed
• Health probes not compatible with the container path

Recommended steps:

• Check container logs in Log Stream → Container Logs.
• Ensure your app listens on the same port as WEBSITES_PORT.
• Add a simple while true; do sleep 30; done to test whether the container stays alive.
• Run the image locally (docker run -p 8080:8080 image) to confirm it behaves as expected.

In almost all cases, the termination time (~4 seconds) indicates the container is exiting by itself, not being forcefully stopped by App Service.