Quiz: SLIs, SLOs and error budgets

Question 10 of 10

One correct answer

The overall 99.5 % SLO over 30 days still holds, while the 99.9 % SLO for /api/ over the same 30 days has gone below zero. How did that happen?

For /api/ the 4xx answers count as bad too.
/api/ has a shorter window, so the outage weighs more.
A counting error — a strict SLO does not break before a soft one.
/api/ has fewer requests and an allowed share five times smaller, and all the 5xx fell on /api/.

Back A–D or ↵ Sign in to answer