17 days ago
Subject: Service not recovering after May 19-20 outage — PostgreSQL connection refused
Description:
My application (app.optimaops.hr) has been down since the major outage on May 19. The PostgreSQL service shows as ACTIVE in the dashboard, but the web service fails to connect on every redeploy with the following error:
sqlalchemy.exc.OperationalError: (psycopg2.OperationalError)
connection to server at "caboose.proxy.rlwy.net" (66.33.22.253), port 35169 failed:
server closed the connection unexpectedly
Steps taken:
Redeployed web service multiple times
Switched DATABASE_URL from internal (postgres.railway.internal) to public URL (caboose.proxy.rlwy.net)
PostgreSQL service shows ACTIVE and "Deployment successful"
Web service still crashes on every attempt
Expected: Web service connects to PostgreSQL and starts normally.
Actual: Connection refused/dropped on every deploy attempt.
Project is on US East region. This has been down since ~22:00 UTC May 19.
Any help appreciated.
1 Replies
17 days ago
This is related to the ongoing service disruption that affected TCP Proxy and US East infrastructure. Your Postgres service's last deployment predates the outage, and we're automatically redeploying workloads we detect as unhealthy, but yours may not have been caught yet. Please try redeploying the Postgres service from its dashboard (it's in your "resource flow-db" project), then once it's healthy, redeploy the web service in "resource flow-app". If it still fails after both redeploys, let us know.
Status changed to Awaiting User Response Railway • 17 days ago
Status changed to Solved dkostovic • 14 days ago