Skip to content
S. Sonbarse
← All labsSimulation

Error budget

Should this incident wake someone up?

Set an SLO, then describe an incident. See the budget burn down, and which multi-window burn-rate alerts fire and when.

30-day SLO
5%
1 h
Monthly budget
43.2 min
Burn rate
50x
Budget used by this incident
6.9%

Wake someone up.

Error budget remaining

100%50%0%Budget exhaustedIncident start1 hPage 14.4xPage 6x
Multi-window alert policies
SeverityBurn rateWindowsFires after
Page14.4x1 h and 5 min18 min
Page6x6 h and 30 min44 min
Ticket1x3 d and 6 hDoes not fire

How the model works

Uses the burn-rate thresholds from the Google SRE Workbook: 14.4x over 1h and 5m, 6x over 6h and 30m, and 1x over 3d and 6h. An alert fires when both of its windows exceed the threshold.

What it leaves out

  • Assumes a constant error rate during one incident and a clean budget before it.
  • Uses a 30-day window and treats every minute as equal, instead of counting requests.

Covers

  • SLOs
  • Burn-rate alerting
  • Incident severity
  • Reliability engineering
Next labWhy was this request denied?