diagnose

A structured debugging approach systematically reproduces, minimizes, hypothesizes, instruments, fixes, and regression-tests hard bugs. It builds a reliable feedback loop to efficiently find and resolve complex system failures or slow performance issues.

To debug a performance regression, follow a disciplined diagnosis loop: build a feedback loop, reproduce the issue, form falsifiable hypotheses, instrument the code, apply a fix, and verify the solution with regression tests.

Creating a reliable reproduction test for an unresponsive application requires building a feedback loop that accurately reflects the issue description. This ensures the debugging loop consistently triggers the system failure for precise instrumentation.

Diagnosing slow performance and system failures requires access to the codebase, execution environment, and debugging tools. This setup enables targeted instrumentation to gather precise information for forming falsifiable hypotheses.

Use hypothesis-driven bug diagnosis when facing hard bugs or performance regressions that resist ad-hoc debugging. It encourages forming falsifiable hypotheses based on observations, ensuring targeted instrumentation and clean regression testing.

After fixing a bug, implement regression tests to verify the solution and conduct a post-mortem. This cleanup ensures a clean, informative bug fix record that prevents future performance regressions and aids future debugging.