kata-incident-triage

What problem does it solve?

It prevents slow, chaotic early incident response by giving an ordered procedure to acknowledge alerts, declare severity, coordinate roles, perform initial diagnosis, and choose mitigation quickly.

Core Features & Use Cases

• First-15-min triage workflow: acknowledge within 5 minutes, declare SEV-1 to SEV-4 objectively, and open the correct war-room/role activation.
• Runbook-driven diagnosis: consult alert runbooks, capture dashboard/log/trace findings with timestamps, and list hypotheses in a testable order.
• Mitigation-first recovery: decide rollback vs forward-fix using deployment/migration/data impact heuristics, apply the mitigation, and verify recovery within 10 minutes.
• Structured communication outputs: produce a channel-ready severity/activations message, a timeline for the IC, and status-page/executive updates when required.
• Use case: during a production latency spike in a service like refund-api, run the triage steps to determine severity, gather early evidence from dashboards/logs/traces, mitigate via rollback or other safe actions, verify recovery, and document a timeline for investigation.

Quick Start

Ask the AI to act as an incident scribe and produce a complete incident triage plan for your next alert using the required steps, outputs, and SEV criteria from the skill.