Question #1169
A company is migrating a web application to Amazon EC2 instances and needs to monitor its performance. The company does not want to trigger alerts for temporary spikes in network latency. However, if network latency exceeds 100ms and HTTP 5xx errors increase simultaneously, immediate action is required. The solutions architect must also ensure false alarms are minimized. What should the solutions architect do to meet these requirements?
Create Amazon CloudWatch composite alarms where possible.
Create Amazon CloudWatch dashboards to visualize the metrics and identify issues manually.
Create Amazon CloudWatch Synthetics canaries to simulate user traffic and detect errors.
Create separate Amazon CloudWatch metric alarms for latency and errors and combine them in a dashboard.
Explanation
Answer A is correct because Amazon CloudWatch composite alarms enable the creation of alarms that trigger only when multiple specified conditions are met simultaneously. In this scenario, a composite alarm would activate only if both network latency exceeds 100ms and HTTP 5xx errors increase, aligning with the requirement to avoid false alarms from temporary spikes in a single metric.
Other options are incorrect because:
- B: CloudWatch dashboards only visualize metrics; they do not automate alerts.
- C: Synthetics canaries simulate traffic but do not directly combine latency and error metrics into a single alarm.
- D: Separate alarms for latency and errors would trigger independently, leading to false alarms if only one condition occurs.
Key points:
1. Composite alarms reduce false positives by requiring multiple conditions.
2. Immediate action is needed only when both latency and errors exceed thresholds.
3. Temporary spikes in one metric (e.g., latency) should not trigger alerts.
Answer
The correct answer is: A