Burn rate
Burn rate is how fast a service is spending its error budget, compared with the pace that would use it up exactly at the end of the window. It is the observed error rate divided by the error rate the service level objective (SLO) allows. A burn rate of 1 is on budget.
How it works
Take a 99.9% SLO over 30 days. The allowed error rate is 0.1%. Suppose 1.44% of requests fail for one hour. The burn rate is 1.44 / 0.1 = 14.4. At that pace the budget lasts 720 hours / 14.4 = 50 hours, and that one hour spends 2% of it.
The Google SRE workbook suggests these alert thresholds for a 30-day window:
Burn rate 14.4 over 1 hour spends 2% of the budget. Page.
Burn rate 6 over 6 hours spends 5%. Page.
Burn rate 1 over 3 days spends 10%. Open a ticket.
Each alert also checks a short window, such as 5 minutes, so it stops firing soon after the problem is fixed.
Why it matters in production
A monthly uptime number arrives too late to act on. Burn rate tells you how urgent a problem is right now. A fast burn wakes someone up, and a slow burn becomes a ticket. This cuts noisy pages and still catches slow leaks.
See also
On the One2N blog
Further reading
