Project 3589ba5e-42d0-40c4-af08-a15c366bb9c7, environment production, region US West (Metal).

Since ~13:00 UTC on June 11 (during the US West incident), private networking inside this project is broken. Evidence:

1. Our main app (izyloc-app, service 90773bbf-c4c7-4a33-ad7f-1caffddba463) gets "connection to postgres.railway.internal port 5432 failed: Connection timed out" on every boot - both IPv6 (fd12:...) and IPv4. DNS resolves correctly; TCP never connects.

2. A second, completely independent app in the same project (evolution-api, with its own evolution-postgres) is also down since the same moment.

3. The Postgres services themselves are healthy ("database system is ready to accept connections" in logs), and the dashboard's own Database panel shows "Attempting to connect..." forever.

4. We restarted AND fully redeployed both the app and Postgres (fresh containers) - DNS resolves to the new container IPs but connections still time out.

5. Workaround proving it's network-only: pointing DATABASE_URL to DATABASE_PUBLIC_URL (public TCP proxy) makes the app work instantly, same code. We are running production through the public proxy right now, which is not acceptable long-term.

A separate project on this same account works fine, so the damage is specific to this project/host. Please restore private networking for this project.