2 months ago
Service mondrian-server in project mondrian (production environment) is online and running, but all requests to the public domain are returning 404 from the edge with x-railway-fallback: true.
Details:
Service: mondrian-server (ID: 0177a737-3569-4862-95b7-2e6fec2681ed)
Project: mondrian (ID: 8813a848-e9c4-49bb-b6d9-6c91b36d21a7)
Environment: production (ID: af1778e4-5c40-4e09-a57f-ab0c8a9590d1)
Latest deployment: f449b168-1151-40b1-b22c-4641d12341e0 (triggered 2026-08-18 14:10:15)
Replicas: 1 running, 0 crashed
App: Python/Gunicorn listening on port 5000
Public domain: mondrian-server-production.up.railway.app
Region: us-west2
Symptoms:
Edge returns HTTP 404 with header x-railway-fallback: true
Edge trace: x-hikari-trace: jfk1.cv40, x-railway-edge: jfk1
Zero HTTP requests reaching the application (deploy logs show app started successfully, but no request logs)
Requests are hitting the edge but failing to route to the backend service
Already tried:
Verified service is online and deployment succeeded
Restarted the deployment
Full redeploy from latest commit
No code changes; issue appeared spontaneously
The app is clearly running and listening on port 5000, but the edge cannot route requests to it. This appears to be a service discovery or edge-to-backend routing issue on Railway's infrastructure.
2 Replies
2 months ago
The deployment completed successfully and the container is running, but the edge is confirmed to be returning its fallback response on the .up.railway.app domain. For services created before a platform migration, this can happen when Private Networking is not yet enabled. Could you check whether Private Networking is enabled in the service's Settings > Networking section? If you see an "Enable Private Networking" button rather than your private network endpoints listed, enabling it and then redeploying should restore routing.
Status changed to Awaiting User Response Railway • about 2 months ago
Railway
The deployment completed successfully and the container is running, but the edge is confirmed to be returning its fallback response on the `.up.railway.app` domain. For services created before a platform migration, this can happen when Private Networking is not yet enabled. Could you check whether Private Networking is enabled in the service's Settings > Networking section? If you see an "Enable Private Networking" button rather than your private network endpoints listed, enabling it and then redeploying should restore routing.
2 months ago
Yes, that fixed it
Status changed to Awaiting Railway Response Railway • about 2 months ago
Status changed to Solved Railway • about 2 months ago