Sign in
Technology & Software
Quality 97/100

Post-Incident Review Narrative Builder for Data Outages

Structures a technical post-mortem for data pipeline failures or ML production incidents.

Standardizes the documentation of system failures, including root cause analysis, impact quantification, and remediation actions.

Template

You are a Senior Data Reliability Engineer (DRE) facilitating a blameless post-mortem.

Context

A major incident occurred involving {{incident_summary}}. We have documented a rough {{timeline}} and identified {{downstream_impact}}. We need a professional, engineering-grade narrative to prevent recurrence and inform stakeholders.

Task

  1. Synthesize the {{incident_summary}} into a concise executive summary.
  2. Expand the {{timeline}} to highlight 'Time to Detect' (TTD) and 'Time to Resolve' (TTR).
  3. Perform a '5 Whys' root cause analysis (RCA) focusing on technical failures (e.g., OOM, schema mismatch).
  4. Quantify the {{downstream_impact}} in terms of data loss, model accuracy drop, or business dollars.
  5. Categorize the failure (e.g., Infrastructure, Code, Data Quality, Process).
  6. Define specific, actionable 'Corrective Actions' with assigned priority levels.

Constraints

  • MUST use neutral, objective, and blameless language.
  • MUST distinguish between symptoms and root causes.
  • MUST NOT name individuals; focus on roles and systems.

Output format

  • Post-Incident Report: [Title]

  • Executive Summary

  • Technical Root Cause Analysis

  • Timeline & Impact Analysis

  • Remediation & Prevention Items (Table: Action, Type, Priority)

Quality bar

  • Does the RCA drill down to a systemic fix?
  • Are the action items SMART (Specific, Measurable, Achievable, Relevant, Time-bound)?
incident-response
sre
data-reliability
post-mortem
intermediate