We can see the deployment failing at Create Container in about 6 seconds with no logs produced. No platform incident or host maintenance event was affecting your service at that time. For image-sourced services in this state, open the service, press Cmd+K (Ctrl+K on Windows/Linux) to open the command palette, and select "Redeploy source image" to pull a fresh copy of the image.

Thanks for the response.

I followed the recommended step exactly: I opened the affected service and used “Redeploy source image” from the command palette to pull a fresh copy of the postgres:16-alpine image.

Unfortunately, the new deployment failed again during the container creation/initialization stage. The service still does not reach application startup.

Could you please investigate the new failed deployment on Railway's side and advise on the next step?

I will not make further configuration changes until we know whether the failure is related to image pulling, container provisioning, networking, or the service configuration.

Thank you.

The deployment's error log shows the OCI runtime cannot find the executable because the start command contains literal triple-backtick characters (markdown code fences) wrapping the actual shell command. The container runtime tries to run a file named with those backticks, which does not exist.

Open the service's settings, find the start command field, and remove the triple backticks from both the beginning and the end. The command should start directly with /bin/sh -c and not be wrapped in any markdown formatting.