← Back to Docs
Recipe: Error budget policy writer
Generate SLO-driven error budget policies from your service metrics.
1
Define your SLO target
Start with a single number: your desired availability percentage. Common targets are 99.9%, 99.95%, or 99.99%. This becomes the anchor for every downstream policy decision.
slo_target: 99.95 # three-nines-five2
Calculate the budget
Error budget = 100% − SLO target. For 99.95%, you get 0.05% allowable error over your measurement window. Translate this into concrete numbers your team can act on.
monthly_budget_seconds = 30 * 24 * 3600 * 0.0005 # 1296s3
Write the policy
Burn rate thresholds determine when to page. A 14.4x burn rate consumes the entire monthly budget in 5 days. Set alerting at 2x (48h exhaustion) for early warning and 10x (6h) for critical.
alert_rules:
- burn_rate: 2x → warning
- burn_rate: 10x → critical
- burn_rate: 14.4x → page