Templates by Incident Class

Different incidents, different templates.

Overview

One postmortem template does not fit every incident. An availability outage demands a tight timeline and customer-impact numbers; a security incident demands disclosure timing and remediation depth; a performance regression demands trend data over time. Per-class templates make each postmortem capture what actually matters for that class instead of forcing every incident through the same shape.

• Different incidents, different templates. Outages, security incidents, performance regressions, near-misses. Each has a different shape.
• Per-class analysis depth. Outage templates emphasise timeline; security templates emphasise disclosure window; performance templates emphasise trend data.
• Outage template. Tight timeline, customer-impact numbers, mitigation-versus-fix split. The shape on-call already expects.
• Security template plus per-class action items. Disclosure timing and remediation depth; per-class action-item format that matches the incident shape.

The approach

Three habits make per-class templates work without becoming bureaucracy: a small set of templates not a sprawling library, quarterly review that prunes what does not work, and onboarding integration so new engineers learn the templates as part of joining.

• Small set of per-class templates. Outage, security, performance, near-miss. Not twelve templates; four to six covers most incidents.
• Documented format with rationale. Per-class the why-this-shape. New engineers inherit the reasoning.
• Quarterly template review. Catches templates that stopped reflecting how incidents actually look. Prune ruthlessly.
• Per-class action items plus onboarding. Action-item format matched to class; per-template the onboarding session covers the why.

Why this compounds

Each per-class template captures a recurring shape so the next incident in that class writes itself faster. The team’s incident maturity deepens; analytics across postmortems become possible because the data is structured the same way every time.

• Learning improves. Right template captures what actually matters. Analysis surfaces faster.
• Operational fit. Postmortem shape matches incident shape. Writing time drops.
• Culture reinforced. Per-class templates signal that incidents are taken seriously enough to differentiate.
• Year-one investment, year-two habit. First template is investment. By the third class, the structure is settled and onboarding fast.