I have two FastAPI backends deployed on Railway in the same project.

Backend A: blog post service (/api/posts) – worked fine before

Backend B: lead-gen service – deployed yesterday

After deploying the lead-gen backend, the blog backend started returning 500 errors on /api/posts.

What I see in logs:

FastAPI/uvicorn is running

Requests reach the route

Logs only show:

"GET /api/posts HTTP/1.1" 500 Internal Server Error

No Python traceback or exception is visible in Railway logs (only access logs)

What I’ve verified:

The route exists and is being hit

The service is running (not a deploy or port issue)

Both services are in the same Railway project

No obvious code changes were made to the blog service

What I suspect:

A shared/project-level environment variable (likely DATABASE_URL) was overwritten when deploying the second service

Or the blog service is now pointing to a different / empty database or schema

Or an exception is being swallowed without logging

Question:
What’s the best way to surface the real Python exception in Railway logs, and how do people usually isolate env-var or DB conflicts when running multiple FastAPI services in the same Railway project?