Hello Railway Support Team,

I am experiencing a critical issue with my PostgreSQL service (Image: ghcr.io/railwayapp-templates/postgres-ssl:16.11).

Context: My environment consists of an Evolution API (WhatsApp bridge) connected to this Postgres instance. Earlier today, the Evolution API started reporting Connection Closed and conflict: replaced errors (Baileys/Socket error). Upon investigation, I found the Postgres service had reached its connection limit (99 active sessions).

Current Status:

1. I have successfully terminated the zombie sessions using pg_terminate_backend.

2. However, the Postgres service is now stuck. When I try to redeploy or restart, it stays in the "Processing" or "Building" state indefinitely without generating new logs.

3. I suspect a Volume Lock or a stuck postmaster.pid file on the persistent storage, but I cannot "Detach" the volume or force a clean boot without risking data loss.

4. I have already corrected the environment variables (DATABASE_URL) and cleared the build cache, but the service remains unresponsive.