BackendDevOps/Cloud
Retry with Exponential Backoff
Also written as Exponential Backoff, Jitter, Retry Storm
Retrying a failed request after progressively longer waits, with randomness added, so a struggling service isn't finished off by everyone retrying at once.
Think of it like
Not everyone redialling the same busy number at the same second.
Junior or senior?
Junior sounds like
Adds retries.
Senior sounds like
Knows naive retries can turn a blip into an outage, and mentions jitter or retry budgets.
Ask them
“Have you seen retries make an incident worse? What happened?”