Use this skill to convert a SLOs, SLIs, SLAs, error budgets, uptime, latency, freshness, recovery objectives, and customer reliability promises question into a concrete artifact with owners, gates, metrics, and recovery paths.
Workflow
- Identify critical user journeys, customers, contractual promises, measurement sources, failure modes, support commitments, and business risk.
- Read
references/service-level-slo-patterns.md. - Separate internal SLO, external SLA, support target, uptime claim, freshness objective, latency objective, and recovery objective.
- Define SLIs, windows, exclusions, error budget policy, alerting, status page behavior, maintenance windows, and customer communications.
- Produce service-level plan, state machine, decision table, event schema, reliability checklist, and promise-review cadence.
When not to use
- Do not use for generic advice the base model already handles without this skill's specific artifact contract.
Guardrails
- Do not promise an SLA if monitoring, incident response, support process, and credit/remedy policy cannot support it.
- Do not measure availability only from server uptime when the user journey depends on dependencies, auth, billing, DNS, or queues.
- Do not hide planned maintenance or partial degradation behind green status.
- Do not treat SLO burn as an engineering-only metric when customer promises or sales claims are affected.
Output format
Service-level context:
Journey / customer promise / measurement source:
SLO and service-level plan:
| Area | Decision | Evidence | Risk | Owner |
| --- | --- | --- | --- | --- |
SLIs, windows, and promises:
- <item> -> <policy, metric, edge case, support note>
Error budget and communication policy:
- <trigger> -> <action, communication, owner>