Service Level Agreement
Our SLA
Public, measurable, and burn-rate alerted. Real-time health at /status.
Service Level Objectives
| Metric | Target | Notes |
|---|---|---|
| Availability target | 99.5% | Monthly. Excludes pre-announced maintenance windows. |
| Error budget per month | 3.6 hours | Burn-rate alerts at 6× and 36× baseline. |
| Cron success rate | ≥ 99% | Per `cron_runs` table; auto-paged on degradation. |
| Read p95 latency | < 500 ms | Public schedule + admin list views. |
| Write p95 latency | < 1000 ms | Booking creation, payment recording, mutations. |
| Recovery point (RPO) | ≤ 5 min | Supabase PITR + daily off-region pg_dump. |
| Recovery time (RTO) | ≤ 4 hours | Quarterly restore drill verifies the target. |
Acknowledgement times
Business hours = Mon–Fri, 09:00–17:00 CET. After-hours P1 events page the on-call rotation.
| Severity | Ack | Resolve target |
|---|---|---|
| P1 — Outage / data loss / security breach | 15 min | Best-effort 24h |
| P2 — Major feature broken for many venues | 30 min | Next business day |
| P3 — Single-venue feature broken | 2 hours | 3 business days |
| P4 — Minor / cosmetic / documentation | 24 hours | Next sprint |
Maintenance windows
- Planned maintenance announced ≥ 72h in advance via the status page
- Default window: Sundays 02:00–04:00 UTC (low-traffic globally)
- Emergency security patches deployed without notice; RFO post-mortem within 5 days
Need to escalate?
P1 incidents — email security@bookingbible.com with subject "P1 — <summary>".