Production Bug
Overview
This playbook covers bugs found in production that are not yet full incidents but still need careful handling.
Why It Matters
Production bugs can affect users silently and recur if the root cause is missed.
Core Concepts
- Reproduce the bug.
- Determine the scope.
- Fix the shared cause if possible.
Mental Models
The reported bug is a clue, not the entire diagnosis.
Best Practices
- Capture user impact.
- Inspect logs and requests.
- Add a regression test after the fix.
Common Mistakes
- Fixing only one visible path.
- Shipping an unverified patch.
- Leaving the bug uncaptured in tracking.
Trade-offs
Small patches are fine when the bug is local, but recurring issues usually justify a broader fix.
Decision Framework
| Scope | Response |
|---|---|
| One route | Local fix may be enough |
| Shared path | Fix the shared code |
| Recurrent bug | Add regression coverage |
Examples
- A broken auth redirect should be fixed in the shared flow, not a single page.
Checklists
- Can I reproduce the bug?
- Is the fix in the shared path?
- Did I add a regression test?
Senior Engineer Notes
Senior engineers care about recurrence. The best production bug fix is the one that stops the next version of the same bug too.