World map

 

BM

World O1: The Control Room

SLIs, SLOs and error budgets

You will learn: how to turn 'the shop works' into a number, a target and a budget for mistakes.

Checked with Prometheus 3 on 1 Nov 2026

The problem

Your boss says: “Checkout must never, ever fail.” Is 100% a good target?

Try it

Divide good checkouts by all checkouts. That share is your SLI (service level indicator): A number that shows how users feel the service, like the share of good checkouts..

Add / kedai_checkout_total after the name, then Run.

Metrics in this data (tap to add):

Time 0 is when the data starts. The query looks back from the eval time.

Predict

Your target, the SLO (service level objective): The target for an SLI over a time window, like 99.9% good checkouts in 30 days., is 99.9% good checkouts over 30 days. The SLI says 99.95%. OK?

Check

A good SLI is something users feel. Swipe right for good SLIs.

Swipe right for good, left for needs work. Or use the buttons.

1 / 5

Share of checkouts that succeed

Watch

The error budget: The bad you may spend: 100% minus the SLO. Spend it on change, not on surprises. is what is left: 0.1% may fail. Over 30 days, how much full downtime is that?

Try it

An outage starts at 40m. Show the SLI only when it misses the SLO.

Add < 0.999 at the end, then Run. Slide between 30m and 60m.

Metrics in this data (tap to add):

Time 0 is when the data starts. The query looks back from the eval time.

You did it!

You can now set an SLI, an SLO and an error budget.

Go deeper (optional)

Spend the budget on purpose: new releases, experiments, planned work. Budget nearly gone? Slow down and fix reliability first. Most teams use a rolling window, like the last 30 days. World O3 turns this into burn-rate alerts.

Real mission

0 of 3 steps done · Your ticks are saved.