Issue Summary

My Flask application (ruml.ai) is experiencing frequent internal server errors (500) due to SSL connection failures when making API calls to Anthropic's Claude API. Multiple users across different locations have reported these errors.

Application Details

• Domain:ruml.ai

• Platform: Railway Pro (recently upgraded from Hobby)

• Framework: Flask with gunicorn

• External API: Anthropic Claude API (multiple calls per user session)

Error Pattern

Consistent SSL connection failures in httpcore library when communicating with Anthropic's servers:

File "/opt/venv/lib/python3.12/site-packages/httpcore/_sync/http11.py", line 217, in _receive_event
data = self._network_stream.read(
File "/opt/venv/lib/python3.12/site-packages/httpcore/_backends/sync.py", line 128, in read
return self._sock.recv(max_bytes)
File "/root/.nix-profile/lib/python3.12/ssl.py", line 1232, in recv
return self.read(buflen)
File "/root/.nix-profile/lib/python3.12/ssl.py", line 1105, in read
return self._sslobj.read(len)
File "/opt/venv/lib/python3.12/site-packages/gunicorn/workers/base.py", line 204, in handle_abort
sys.exit(1)
SystemExit: 1

Troubleshooting Already Attempted

1. Upgraded to Railway Pro - Errors persist after upgrade

2. Added Anthropic client timeout configuration:

   client = anthropic.Anthropic(    api_key=app.config['ANTHROPIC_API_KEY'],    timeout=httpx.Timeout(        timeout=60.0,        connect=10.0,        read=30.0    ),    max_retries=2)
   

3. Application-level error handling - All API calls have try/catch blocks, but errors occur at SSL level before application can handle them

Additional Context

• Errors occur across different user sessions and geographic locations

• Same application works perfectly in local development

• Also seeing some gunicorn HTTP parsing errors (separate issue)

• Application makes 4-5 sequential API calls per user session

Questions for Railway Support

1. Are there known network connectivity issues between Railway and Anthropic's API endpoints?

2. Could this be related to Railway's SSL/TLS configuration or network routing?

3. Are there recommended configurations for applications making frequent external API calls?

4. Should I consider different hosting regions or network configurations?

Current Impact

Users are experiencing failed sessions and 500 errors, preventing proper product validation and user feedback collection.

Any guidance on Railway-specific network configuration or known issues with external API connectivity would be greatly appreciated.