Hi Railway Team,

I'm experiencing a critical DNS resolution issue that is blocking my production application from connecting to Supabase.

PROBLEM SUMMARY:

My service cannot resolve the hostname "rwxhwumfsoyzjizvqxgp.supabase.co"

Error: [Errno -2] Name or service not known

SERVICE DETAILS:

- Service Name: videosync-processor

- URL: https://videosync-processor-production.up.railway.app

- Plan: Hobby (8 GB RAM, 8 vCPU)

- GitHub Repo: Multi-Kamera/videosync-processor

- Branch: main

TIMELINE:

1. November 20, 2025: Service worked perfectly, Supabase connection OK

2. November 21, 2025 16:30 CET: Deleted and recreated service from same GitHub repo (to force redeploy)

3. November 21, 2025 17:00 CET: DNS Error appeared

4. November 21, 2025 19:15 CET: Deleted service again, waited 10 minutes, recreated - DNS Error persists

CURRENT STATUS:

Health check works: https://videosync-processor-production.up.railway.app/

Service is Active and running

Environment variables correctly set (SUPABASE_URL, SUPABASE_SERVICE_ROLE_KEY)

Same code that worked before November 21, 16:30

Any attempt to connect to Supabase fails with DNS error

ERROR DETAILS:

When service tries to connect to Supabase:

httpx.ConnectError: [Errno -2] Name or service not known Host: rwxhwumfsoyzjizvqxgp.supabase.co

TROUBLESHOOTING ALREADY DONE:

1. Verified environment variables are correct

2. Tested with Python httpx, requests, and socket libraries - all fail

3. Service Reset (twice) - problem persists

4. Confirmed same code worked before service recreation

5. /etc/resolv.conf exists but DNS resolution fails at system level

TECHNICAL ANALYSIS:

- Python socket.gethostbyname() fails to resolve the hostname

- This suggests the Railway container cannot reach DNS servers

- The issue appeared immediately after service recreation

- Old service instance (before 16:30 CET) worked perfectly

- New service instances (after 16:30 and after 19:15) both have same DNS issue

QUESTION:

Can you either:

1. Migrate my service to infrastructure configuration that worked before November 21, 16:30?

2. Fix DNS resolution for rwxhwumfsoyzjizvqxgp.supabase.co on my current service?

3. Provide guidance on network configuration that might resolve this?

The service worked perfectly with the exact same setup until the service recreation. This suggests the issue is related to which physical infrastructure/network configuration the new service instance was assigned to.

IMPACT:

- Production application is blocked

- Cannot process video merging (core functionality)

- Cannot connect to database

I appreciate any help you can provide. Happy to provide additional logs or diagnostics if needed.

Best regards,

Tammo

Additional Technical Details:

DOCKERFILE:

- Based on python:3.11-slim

- Uses Gunicorn with --timeout 600

- FFmpeg installed via apt-get

- No custom DNS configuration

PYTHON DEPENDENCIES:

- httpx (for HTTP requests)

- supabase-py (Supabase client)

- Flask/Gunicorn (API server)

DNS TEST RESULTS:

- socket.gethostbyname('rwxhwumfsoyzjizvqxgp.supabase.co') → Name or service not known

- nslookup/dig not available in container

- /etc/resolv.conf readable but DNS resolution fails

The old service instance (deleted on Nov 21, 16:30) successfully resolved this exact same hostname for weeks.