No circuit breaker on critical path

What it is

A dependency on a critical path has no circuit breaker: every request keeps calling it at full rate regardless of how many recent calls failed.

Why it matters

When the dependency is down, each request waits for its own timeout before failing. Your latency becomes their timeout, workers pile up, and the failure spreads to parts of your app that do not use that dependency at all.

How to fix it

Wrap the call in a breaker that opens after a failure threshold, fails fast while open, and probes periodically to close. Pair it with a fallback so an open breaker degrades a feature rather than the whole request.