incident-triage-and-diagnostics

name	incident-triage-and-diagnostics
description	Triage reliability incidents by isolating likely fault domains and defining safe mitigation and recovery checks. Use during outages, regressions, failed deployments, or recurring degradation.
argument-hint	Describe the symptoms, timeframe, impacted flows, recent changes, and the logs or telemetry already available.

Incident Triage and Diagnostics

When to Use

• Active reliability incident or recurring degradation
• Failed deployment or sudden change in error/latency patterns
• Need fast isolation of likely fault domains

Quick Checklist

1. Define impact, blast radius, and timeframe.
2. Correlate symptoms across logs, metrics, and traces.
3. Isolate probable fault domain and confidence level.
4. Recommend immediate mitigation and verification steps.

Deep Workflow

1. Capture incident scope: impacted users/flows, first-seen time, severity, and current status.
2. Establish a timeline using key telemetry markers (error spikes, latency shifts, deploy/config events).
3. Segment by likely domains (frontend, backend API, dependency, configuration, infrastructure).
4. Test the top hypothesis with focused evidence queries and minimal reproducible checks.
5. Define short-term mitigation (rollback, feature flag, throttling, config correction) with tradeoffs.
6. Define validation checks to confirm recovery and prevent false positives.

Decision Points

• If evidence is mixed, prioritize mitigations with lowest blast radius first.
• If root cause confidence is low, separate mitigation from root-cause claims.