Incident Response
In one line
Incident response is the structured process of handling outages — detect, contain, resolve, and learn from incidents.
In simple words
Incident response follows a structured lifecycle to minimize impact and restore service quickly.
Phases: 1. Detection — alerts trigger 2. Triage — assess severity, assign Incident Commander 3. Investigation — identify root cause 4. Containment — stop the bleeding 5. Resolution — fix and restore 6. Postmortem — blameless analysis
Severity: SEV1 (total outage), SEV2 (major), SEV3 (minor), SEV4 (cosmetic).
Example
1. DETECTION: Alert fires, on-call paged
2. TRIAGE: Acknowledge, declare SEV2, assign IC
3. INVESTIGATION: Check dashboards, traces → DB pool exhausted
4. CONTAINMENT: kubectl rollout undo deployment/api
5. RESOLUTION: Error rate drops to 0%, monitor 15min
6. POSTMORTEM: Write blameless analysis, action itemsStructured process: detect → triage → investigate → contain → resolve → postmortem. Each phase has clear actions. Goal is fast resolution and learning.
Common confusions
Confused with: Runbook
The difference: A runbook is a pre-written procedure for specific scenarios. Incident response is the overall process of managing an incident. You follow runbooks during incident response.
Related terms
Related Resources
Related Lessons
- Incident Management
Incident response.
Decode error messages
Plain-English explanations of common errors — what they mean, why they happen, and how to fix them.