Back to Docs
Recipe

Error Budget

Define how much failure your system can absorb before you must stop shipping features and invest in reliability.

What It Is

An error budget is the inverse of your SLO. If you promise 99.9% uptime, your error budget is 0.1% — roughly 43 minutes of acceptable downtime per month. When the budget is exhausted, all feature work freezes until reliability is restored.

The Formula

error_budget = 1 − SLO_target

Example: 99.95% SLO → 0.05% budget → ~22 min/month

Burn Rate Alerts

  • 2% in 1 hour — page the on-call engineer
  • 5% in 6 hours — escalate to incident lead
  • 10% in 3 days — freeze feature deploys

When Budget Hits Zero

Stop all non-critical deploys. Divert engineering capacity to postmortems, automated testing, and observability gaps. Resume feature work only after the budget regenerates or the SLO window resets.

Meridian tip: Track error budgets per-service, not just globally. A single noisy endpoint can burn the whole team's budget before anyone notices.