- An SLO is a target for an SLI over a window: "availability of at least 99.5 % over 30 days"; the window is rolling.
- Error budget = (1 − SLO) × requests in the window; left = budget − bad answers; while more than zero is left, the SLO holds.
- Burn rate = share of bad answers over a short window / (1 − SLO): at 1 the budget runs out at the end of the window, at 180 — in 4 hours.
- A stricter SLO for part of the requests has a smaller budget, and the same outage may overspend it.
- Yesterday's outage ate about two thirds of the budget, and together with the errors of the other days about fifty are left in reserve: the 99.5 % SLO still holds, but a failed rollout today may break the target.
Step 6 of 6 · summary · 3 min
Summary
Found an inaccuracy? Report an error in the lesson