Following the May 19 2026 GCP outage, my Postgres service shows ACTIVE / Online after dashboard Redeploy ~30 mins ago, but external connections via the public proxy still fail with P1001.

Unlike the standard catatonit pid1 crash-loop pattern, my container is reporting healthy — but the service is not actually serving connections. Suspected stale/inconsistent volume state from the outage that the container can mount but Postgres cannot serve from.

Symptoms:

• Postgres deployment shows ACTIVE in dashboard, Activity log: "Redeployment successful"
• TCP path is open: Test-NetConnection to shuttle.proxy.rlwy.net:14172 returns TcpTestSucceeded: True
• Every Prisma connection attempt returns P1001 "Can't reach database server"
• DATABASE_URL in Variables tab matches .env.local exactly
• Tried: 1x Restart, 1x Redeploy, connect_timeout=15 appended — no change
• Catatonit pid1 errors appeared in earlier failed restart attempts overnight but not in current ACTIVE deployment

Per the recovery FAQ, requesting volume relocation to a healthy node. Container preserves intact, please do not reprovision empty — data is critical.

Project ID: 006ce096-1681-4034-b162-491b5212800f

Service ID: 95581856-8719-451b-9227-65afec8b4cfd

Volume: postgres-volume

Image: ghcr.io/railwayapp-templates/postgres-ssl:18

Last known good operation: 2026-05-19 22:33 BST (Postgres checkpoint log line)

Outage window: 22:20 UTC May 19 - 06:14 UTC May 20

Thanks for working through this.