concepts/

wiki/concepts/{concept-slug}.md

Definitions of the terms the wiki keeps using — SEV1 vs SEV2, error budget, MTTR, paging policy. Shared vocabulary so two pages mean the same thing by the same word.

What a page looks like

---
type: concept
name: SEV2
status: reviewed
tags: [severity, paging]
last_updated: 2026-04-26
---

## Definition
A SEV2 is a partial outage affecting a single tier-1 service or
a degraded experience for >5% of users for >15 minutes …

## When to declare
- p95 latency on any tier-1 service > 2s for 5+ minutes
- error rate > 1% sustained
- a customer-facing flow is broken but a workaround exists

## Paging
Pages the service oncall immediately. Escalates to team lead
if not acknowledged in 10 minutes. No leadership page.

## Related
- [[sev1]] — full outage / data loss
- [[sev3]] — minor degradation
- [[paging-policy]]

Sections — flexible, but typical

1. Definition — what the term means, one paragraph.
2. When to use / declare — concrete triggers.
3. Related — links to nearby concepts and the policies that use this one.
4. Notes — open questions, contradictions, edge cases.

Rules

1. File name is the concept slug in kebab-case: sev2.md, error-budget.md, paging-policy.md.
2. Concepts have no type-specific frontmatter — just the universal block.
3. Reuse over reinvent — before creating a new concept, search the existing ones for an equivalent name.
4. If a concept page contradicts how the term is used on an incident page, log it as a contradiction; don't silently revise either side.