hard20 pts · first answer countsdistributed-systemsresilience

Outage that outlives its trigger

A 2-minute database failover caused errors. The database is healthy again, but the API stays overloaded for an hour, with most work spent on retries of requests whose clients already gave up. What is this, and what breaks the loop?

A.A memory leak; restart the API pods one by oneB.A metastable failure; shed load or cut retries so the system can drain, and drop work whose deadline has passedC.A DNS cache problem; flush resolversD.Normal behaviour; wait for the error budget to reset

Discussion 0

Join free to post in the discussion.

Outage that outlives its trigger

Discussion 0

Related questions