investigate

Hypothesis-driven root cause analysis formulates multiple theories and tests them against logs, source code, and historical RCAs. It consolidates evidence from repository artifacts and external documentation to converge on a validated root cause for confident remediation.

Investigating intermittent issues correlates request logs, recent deploys, and third-party API docs with past postmortems. This evidence-gathering process validates scope and impact to identify the primary cause and recommend targeted fixes.

Start an investigation by providing the incident or finding reference to trigger the workflow. It gathers evidence, validates hypotheses against historical data, and produces a structured timestamped report saved to docs/investigations/.

It supports both incident response and proactive findings. For proactive findings, it runs in confirmation and triage mode to validate scope and impact against past RCAs and external documentation before full escalation.

Hypothesis testing searches repository artifacts, reviews past investigations, and consults external documentation. While it can analyze available logs and code, historical RCAs and external docs provide critical validation for scope and impact assessment.

Use a structured investigation workflow when you need to consolidate evidence from logs, code, and docs to make confident remediation decisions. It prevents guesswork by following a hypothesis-driven RCA process that validates impact against historical data.