What can break in production, and how would I debug it?
Understand the concept
Production failures often come from conditions absent in small examples: parallel requests, large datasets, expired credentials or unavailable providers. A review should consider how the code exposes and recovers from these conditions. Observability is part of the design because a hidden failure is harder to correct safely.
Important ideas
-
Check scale, parallel requests, stale data and unavailable dependencies.
-
Use request IDs, logs, metrics and traces to locate the failing step.
-
Reproduce the failure and add a focused regression test.
-
Check whether logs contain stable request or operation IDs and whether sensitive inputs are excluded.
Can you explain it without the notes?
What can break in production, and how would I debug it?
Use your own words. Explain an example and one limitation before checking the explanation.
Check your explanation
Production failures often come from conditions absent in small examples: parallel requests, large datasets, expired credentials or unavailable providers. A review should consider how the code exposes and recovers from these conditions. Observability is part of the design because a hidden failure is harder to correct safely.
Turn this concept into a clear answer.
- Explain how you would distinguish the suspected defect from other causes once the code is running.
Adapted from 11_AI_CODE_REVIEW.md · “What can break in production, and how would I debug it?”. Examples are study fragments; framework behavior can depend on your version.