Project ID: 31ea39d8-9878-471a-9f56-327a6a199fc6

Environment: production

PostgreSQL Service ID: f0ea6889-d7ca-4754-a60a-e92bbedb2be4

Issue: PostgreSQL private networking (postgres.railway.internal:5432) is not accepting connections, despite the container showing as running with normal CPU/memory metrics (~400MB memory usage).

Symptoms:

• All application services fail to connect with: connection to server at "postgres.railway.internal" (fd12:2ebd:b579:1:2000:b2:eae8:d41f), port 5432 failed: Connection refused

• pg_isready returns "no response" for the private networking address

• PostgreSQL deployment status shows SUCCESS

• Container metrics show ~400MB memory, active CPU usage

• Volume is healthy (11GB/50GB used)

Timeline:

• ~16:23 UTC (2026-02-16): PostgreSQL container was stopped (logs show "Stopping Container") after a flood of duplicate key errors

• Multiple redeploys and restarts attempted via CLI - all show SUCCESS but connections still fail

• Private networking appears to be broken while the container itself is running

What we've tried:

• railway redeploy --service Postgres (SUCCESS, no change)

• railway restart --service Postgres (no change)

• Redeploying dependent services (all crash on DB connection)

Configuration:

• Image: ghcr.io/railwayapp-templates/postgres-ssl:17

• Region: US West (California)

• Volume: 50GB at /var/lib/postgresql/data

• Private networking DNS: postgres.railway.internal

Request: Please help figure out why private networking is not routing to the PostgreSQL container. The container appears to be running but is not reachable via the internal DNS/IP addresses.