← Technical notes

Build an Incident Timeline Before Chasing a Production Fix

A practical, non-client-specific framework for reconstructing what changed, what failed, and what evidence to preserve.

Abstract incident trace with a highlighted decision point

Start with the sequence

Before proposing a fix, write down what was observed, when it was observed, and what changed around it. A shared sequence is more useful than a collection of guesses.

Preserve the useful evidence

Capture request identifiers, relevant configuration context, and the exact path being discussed. Keep the record focused on facts that can be checked again.

Choose the next test

End the timeline with the smallest safe question that will reduce uncertainty. This framework is general guidance, not a report of any client incident.