observability-slo
Installation
SKILL.md
Observability and SLO (lite)
Three pillars (minimum viable)
| Pillar | Start here |
|---|---|
| Logs | Structured JSON; request/job ID; severity |
| Metrics | RED/USE for services; duration, errors, throughput |
| Traces | One trace per request/job across critical hops |
SLI -> SLO flow
- Pick SLI — measurable user- or business-visible signal.
- Set SLO — target over window (e.g. 99.9% availability / 30d).
- Error budget — 100% - SLO; spend triggers policy (slow features, freeze risky deploys).
- Alert on budget burn — fast burn (pages) vs slow burn (ticket).