3 steps · SLOs & reliability
Review a reliability budget
Calculate the allowance, inspect burn rate and connect the result to your release policy.
BringAn agreed SLO, eligible event or time totals, bad outcomes and a policy owner.
Finish withAn error-budget report, a two-window burn-rate check and a reviewed policy checkpoint.
Step 1
Calculate the allowance
Choose request-based or time-based availability, then supply the SLO and matching eligible and bad totals. The result shows how much of that population is allowed to fail and how much allowance remains.
Before moving on: The numerator and denominator need the same service scope and measurement window. Note excluded traffic and missing telemetry so the report explains what its allowance covers.
Save the report. A time-based allowance and a request-based allowance are different definitions.
Step 2
Inspect the rate of consumption
For an event-based SLI, the long window describes sustained consumption and the short window shows recent errors. Enter their separate counts with the same indicator and a common end time.
Before moving on: The report compares each rate with the chosen threshold. The short window is part of the long window, so adding their consumption would count some events twice.
Enter the window counts separately. The calculator does not import the previous budget report or convert minutes into requests.
Step 3
Record the release-policy decision
Return to the budget draft and enter your team’s consumption checkpoint, the action it triggers and the decision owner. The report keeps that policy beside the allowance it applies to.
Before moving on: The decision owner needs the measurement scope, remaining allowance and any applicable exception to determine whether the next change can proceed.
Export the report for review. A passing calculation does not authorize a release.