Question

A cloud operations team receives complaints that a payment API intermittently returns HTTP 500 errors, but the existing CPU and memory alarms never trigger during these events. Application logs are already being shipped to the cloud provider's centralized logging service. The team wants automated notification the moment the 500 error rate crosses a defined threshold, without deploying additional agents or code changes. Which approach best meets this requirement?

Accepted Answer

Create a log-based metric filter that counts HTTP 500 entries in the ingested logs and attach an alarm that notifies the team when the count exceeds the threshold over a set period.

Answer

Increase the CPU utilization alarm sensitivity so it triggers earlier and correlates with the API errors.

Answer

Enable detailed instance monitoring at 1-minute granularity to capture the error spikes in the hypervisor metrics.

Answer

Configure a synthetic canary to load-test the endpoint continuously so failures generate alerts.