Menu
Define SLIs, SLOs, and implement burn-rate alerting
Installer avec Codex ou Claude Copiez ce prompt, collez-le dans Codex, Claude ou un autre assistant, puis laissez-le vérifier la page du skill et l'installer pour vous.
Basé sur la classification professionnelle SOC
Track cache hit rates, latency, and detect cache-related issues
Instrument database queries, connection pools, and detect N+1 queries
Capture errors with rich context for debugging and alerting
Implement liveness, readiness, and dependency health checks
Plan backend observability using RED + USE + 4 Golden Signals + JTBD
Monitor message queues, job processing, and detect backpressure
| name | slo-alerting |
| description | Define SLIs, SLOs, and implement burn-rate alerting |
| triggers | ["SLO definition","SLI metrics","error budget","burn rate alerting"] |
| priority | 2 |
Define SLIs, set SLO targets, alert on burn rate (not raw error rate).
| Term | Definition | Example |
|---|---|---|
| SLI | Quantitative measure | % successful requests |
| SLO | Target for SLI | 99.9% success |
| Error Budget | Allowed failure | 0.1% = 43 min/month |
| Burn Rate | Budget consumption speed | 10x = exhausted in 3 days |
Availability: successful_requests / total_requests
Latency: requests_under_threshold / total_requests
Error Rate: error_requests / total_requests
Alert on how fast you're consuming budget, not raw error rate:
| Alert Level | Burn Rate | Time to Exhaust |
|---|---|---|
| Page (critical) | 14.4x | 2 days |
| Page (warning) | 6x | 5 days |
| Ticket (medium) | 3x | 10 days |
Use long + short windows to balance speed and noise:
# Critical: Fast burn (14.4x over 1h AND 5m)
- alert: HighBurnRate_Critical
expr: (rate_1h / budget > 14.4) and (rate_5m / budget > 14.4)
severity: critical
# Warning: Slower burn (6x over 6h AND 30m)
- alert: HighBurnRate_Warning
expr: (rate_6h / budget > 6) and (rate_30m / budget > 6)
severity: warning
references/methodology/sli-slo-framework.md