Incidents
A deployment just raised the error rate to 20%. What do you do?
Whether the candidate reduces user impact before seeking understanding, and whether they can run the response as a coordinated activity rather than as a debugging session.
The situation behind the question
A rolling deployment is 30% complete. The error rate has gone from 0.2% to 20%, concentrated on one endpoint. The on-call engineer joined the team two weeks ago and has never rolled back this service.