DOP-C02 Question 244
Single answerYour DevOps team manages a highly dynamic web application hosted on Amazon EC2 instances. The traffic pattern for the application is unpredictable, with frequent spikes and drops. You want to monitor the CPU utilization of the EC2 instances and receive alerts only when the usage deviates significantly from typical patterns, without generating false alarms during normal traffic fluctuations. Which approach would best meet this requirement?
- A
Create a standard Amazon CloudWatch alarm with a static threshold for CPU utilization.
- B
Enable Amazon CloudWatch anomaly detection for the CPU utilization metric and configure an alarm based on the anomaly detection band.
- C
Use AWS Auto Scaling to dynamically adjust the number of EC2 instances based on CPU utilization thresholds.
- D
Set up a CloudWatch Logs Insights query to analyze CPU utilization trends and manually monitor deviations.
Show answer and explanation
Correct answer: B
Explanation
Amazon CloudWatch anomaly detection is designed to dynamically learn the typical behavior of a metric, such as EC2 instance CPU utilization. By enabling anomaly detection and configuring an alarm based on the anomaly detection band, you can minimize false alarms caused by normal fluctuations and focus on significant deviations from the expected pattern. This approach is more effective for highly dynamic workloads compared to static thresholds or manual monitoring.
- A. Incorrect.
This option does not account for dynamic traffic patterns. A static threshold could lead to false alarms during normal fluctuations or miss actual anomalies when the baseline changes.
- B. Correct.
This is the correct approach. CloudWatch anomaly detection learns the typical behavior of a metric and creates a band of expected values. An alarm can then notify you only when the metric deviates significantly from the learned pattern.
- C. Incorrect.
While Auto Scaling can adjust resources based on thresholds, it does not directly address the need for anomaly detection or monitoring unexpected deviations in metric behavior.
- D. Incorrect.
Manually monitoring CPU utilization trends is not scalable or efficient for detecting anomalies in real time, especially for a highly dynamic application.