How much can break?

An SLOService Level Objective — a target for how reliable a service should be, like 99.9% of requests served OK over a month. is a promise; the error budgetThe flip side of the SLO: the small fraction you're allowed to fail. Spend it on risky deploys; run out and you've broken the promise. is the slack it leaves you. Pick a target and see exactly how much downtime — or how many failed requests — you're allowed before you've broken it, and how fast you're spending it.

Target
Keep% available over awindow.
That's an error budget of 0.1% — up to 43m 12s of downtime per 30 days.
Allowed downtime
Per periodDowntime
day
1m 26s
week
10m 5s
30 dayswindow
43m 12s
quarter
2h 10m
year
8h 46m
Budget burnBurn rate: how fast you're spending the budget relative to time. 1× lands you at empty exactly on reset day; 2× burns it in half the window.
Window50% in
So far:of downtime.
31m 12sBudget left
27.78%Used
0.56×Burn rate
On trackOn track at 0.56× the budgeted pace, with 31m 12s to spare.
based on the SRE error-budget model