hard20 pts · first answer countsreliabilitydistributed-systemsperformance

Work continues after the client gives up

Mobile clients time out after 2 seconds. Your API calls a downstream service with a 10-second timeout. When the downstream slows down, the API keeps working on thousands of requests nobody is waiting for and runs out of resources. What is the best fix?

A.Propagate a deadline: each hop uses the remaining time budget as its timeout and cancels work when it runs outB.Raise the mobile client timeout to 10 secondsC.Remove the downstream timeout so calls always completeD.Add three automatic retries in the API so slow downstream calls get another chance before the client gives up

← Stale value returns after cache delete Trusting a Bloom filter answer →