The Problem:

• Your Postgres container is crashing because the Railway volume isn't being mounted at runtime, even though it's correctly configured in your service settings
• The volume exists and contains ~1.06 GB of data, but the container receives an empty mount path instead of /var/lib/postgresql/data
• This triggered a crash-restart loop (12 deployments in ~17 seconds), which left the current deployment stuck in a pending state

VolumeCreate via GraphQL succeeds, mountPath is /var/lib/postgresql/data, but the postgres container reports an empty mount path at runtime.