root-cause-analysis

Root cause analysis traces failures backward through causality—Failure, Trigger, Mechanism, Enabling condition, Root cause—to identify the deepest controllable factor, supported by evidence like logs, traces, or test results, ensuring you fix the true origin and prevent recurrence.

Document your 5 whys analysis by defining the failure, detailing reproduction steps, mapping the why-why causal chain, identifying the root cause, proposing three fixes, and listing verification requirements to ensure the incident is fully resolved and documented.

Perform root cause analysis on test failures and performance regressions by moving backward through causality from the observed failure to its root cause, backing every causal claim with concrete evidence such as code references, diffs, logs, or test results.

Reproduction steps are a core part of the structured output, but the analysis can still proceed by moving backward through causality and supporting causal claims with available evidence like logs, traces, and code references to find the root cause.

Tracing the enabling condition is necessary because the trigger is just the immediate event; the deepest controllable and fixable root cause lies further back in the causal chain, and identifying it allows you to implement fixes that prevent recurrence.

A complete root cause analysis document includes the failure definition, reproduction steps, a why-why analysis tracing the causal chain, the identified root cause, three preventive fixes, and the verification requirements needed to confirm the resolution.