We’re running into persistent 500 errors in production on Railway and trying to understand whether this is a Redis resource or networking issue.

Setup

• Web: Django + Gunicorn

• Redis: used via cacheops

• PostgreSQL: separate container

• All services deployed separately on Railway

What’s happening

When accessing our endpoint, the app crashes with 500.

From logs:

1. Django template rendering calls a model property (active_scenario)

2. That triggers a queryset wrapped with cacheops

3. Redis call times out:

   RuntimeWarning: The cacheops cache timed out!
   Error: Timeout reading from socket
   

4. Gunicorn worker times out:

   WORKER TIMEOUT
   Worker was sent SIGKILL
   

5. Endpoint returns 500

What we tried

• Restarted Redis service

• Restarted Web service

• Issue persists even after restart

• PostgreSQL shows no critical errors

Questions

• Has anyone seen Redis Timeout reading from socket issues on Railway recently?

• Could this be caused by Redis RAM/CPU limits?

• Is there a way to inspect Redis memory usage or connection counts on Railway?

• Could this be an internal networking latency issue between services?

This is impacting our production admin interface, so any guidance would be greatly appreciated.

Thanks!