Data incidents require structured response:
Detect: Alert fires. Acknowledge within SLA.
Triage: Assess impact. Which downstream systems affected? Is this critical?
Communicate: Notify stakeholders. "Revenue dashboard may show incorrect data."
Investigate: Use lineage to trace root cause. Check recent changes.
Fix: Correct the issue. Backfill affected data if needed.
Postmortem: Document what happened, why, and how to prevent recurrence.
Prepare runbooks for common scenarios. Panic debugging at AM wastes time.