SLO Compliance Calculator
Calculate actual SLO compliance from total and failed requests, and the burn rate against your target.
Inputs
Total requests served in the measurement window.
Requests that counted as SLI failures (errors, timeouts, SLA breaches).
The committed service level objective.
Actual SLO
99.8800%
Burn Rate
1.20×
SLO Compliant
false
Margin vs Target
-0.0200pp
Allowed Failed Requests
1,000requests
Step by step
Successful requests
1000000 − 1200
= 998,800
Actual SLO: successful ÷ total
998,800 ÷ 1,000,000
= 99.8800%
Burn rate: actual error rate ÷ allowed error rate
0.1200% ÷ 0.1000%
= 1.20×
How it works
SLO compliance measures how your actual success rate compares to a committed target: actual_slo = (total − failed) ÷ total × 100. Burn rate expresses this as a multiple of your allowed error budget consumption rate — a burn rate of 1x means you're consuming your error budget exactly on pace to exhaust it by the end of the period; a burn rate of 10x means you'd exhaust a 30-day budget in 3 days.
Formulas
Actual SLO
actualSLO = ((totalRequests - failedRequests) / totalRequests) × 100
- T
- Total requests in the measurement window
- F
- Failed requests
- S_a
- Actual SLO percentage
Burn rate
burnRate = actualErrorRate / allowedErrorRate
- e_a
- Actual error rate percentage
- e_t
- Allowed error rate (100 - SLO target)%
- B
- Burn rate (1× = on pace, >1× = over budget)
Frequently Asked Questions
What does a burn rate greater than 1 mean?
It means your current error rate exceeds what's sustainable to stay within the SLO for the full measurement period — if it continues at this rate, you'll breach your error budget before the period ends.
How is burn rate different from actual SLO percentage?
Actual SLO percentage is a point-in-time snapshot of overall compliance; burn rate expresses velocity — how fast you're using up the allowed error budget relative to a sustainable pace.
Should I alert on SLO percentage or burn rate?
Most mature SRE practices alert on burn rate (e.g. multi-window multi-burn-rate alerts) because it catches fast-developing problems early, well before the cumulative SLO percentage itself would breach.
What counts as a 'failed' request for SLI purposes?
This depends on your SLI definition — commonly 5xx errors, requests exceeding a latency threshold, or business-logic failures. Be consistent about the definition across your SLO reporting.