medium10 pts · first answer countsreliabilitydistributed-systems

Retrying after a shared outage

A downstream service goes down briefly. 1,000 clients each retry every 1 second exactly, and the service falls over again each time it comes back. What retry policy is better?

A.Retry immediately in a tight loop until it worksB.Retry exactly every 2 seconds instead of every second, so the service gets twice as long to recoverC.Never retry; return an error at onceD.Exponential backoff with random jitter and a limited number of attempts

← Liveness probe restarts every pod What a circuit breaker does →