All quizzesHard
Reliability & Availability — Series 2
Preview — 3 of 10 questions
An API has a 500 ms deadline and has already spent 350 ms before calling a dependency. What timeout policy respects the remaining budget?
ARetry indefinitely while keeping the original client socket open.
BUse a 1-second timeout because the dependency has a separate SLO.
CPass a deadline with at most 150 ms remaining, allowing response overhead.
DGive the dependency a fresh 500 ms timeout on every attempt.
Peak load needs four equal instances. Five are deployed. Which failure can this capacity plan cover at unchanged load, ignoring other bottlenecks?
AOne instance failure while retaining the four-instance requirement.
BAn arbitrary traffic increase because one spare absorbs every burst.
CTwo simultaneous instance failures without reducing admitted work.
DLoss of the entire zone containing all five instances.
A recovery plan allows at most 20 minutes of outage and at most 2 minutes of lost committed data. What are the objectives?
ABoth are RPO because both concern a database.
BBoth are RTO because both use time units.
CRTO 20 minutes; RPO 2 minutes.
DRTO 2 minutes; RPO 20 minutes.
Sign up free to play
Answer all 10 questions (7 more), see explanations for every answer, and track your score.