Which condition is the most appropriate subject for a Smart Alert?
Select an answer to reveal the explanation.
Short Explanation
You don't alert on one-off blips. Sustained rise in error rate or latency on user-dependent service is meaningful and actionable. One slow request is weather; sustained shift is problem.
Full Explanation
The correct condition is the sustained rise in error rate or latency on a service that users depend on. The sustained change in error rate or latency on a user-facing service is both meaningful and actionable, which is what an alert should be. The completion of each individual trace through the system is incorrect because individual trace completion is normal operation, not an exceptional condition. The creation of a new dashboard by a team member is incorrect because dashboard creation is a user activity with no operational impact. The single request that took slightly longer than the previous one is incorrect because single-request variation is expected noise; alerting on it guarantees constant firing without indicating anything. For the exam, look for sustained, impactful and actionable when identifying a valid alert condition.