QuestionQ41

VerifiedImplementing observability practices and troubleshooting issues

You are responsible for supporting a web application running on Compute Engine that provides a booking service to thousands of users. Shortly after a new feature is released, your monitoring dashboard shows that all users are experiencing latency at login. You want to mitigate the impact of this incident on your users. What should you do first?

Explanation

When a service disruption occurs immediately after a new release, the fastest and safest mitigation is to roll back that release, since it is the most probable cause of the regression and reverting it restores the previously known-good, stable state for all affected users. This follows standard incident management and SRE practice of prioritizing rapid mitigation of user impact over root-cause investigation or further changes during an active incident. Reviewing monitoring data or deploying additional changes (including a new release intended as a fix) does not immediately stop the user-facing impact and can prolong or worsen the outage, while resizing VMs addresses a capacity assumption that isn't confirmed as the actual cause.

Learn more

Community Discussion

No comments yet. Be the first to start the discussion!