SOA-C02 Question 18
Single answerAn organization is running an auto-scaling group of EC2 instances behind an Application Load Balancer (ALB). They want to ensure that their operations team is notified if the average CPU utilization of the instances in the auto-scaling group exceeds 70% for 5 consecutive minutes. How can this be achieved using Amazon CloudWatch?
- A
Create a CloudWatch alarm on the average CPU utilization metric of the auto-scaling group, set the threshold to 70%, and configure the alarm to trigger an Amazon SNS notification.
- B
Create a CloudWatch alarm on the CPU utilization metric for each individual EC2 instance, set the threshold to 70%, and configure the alarm to trigger an Amazon SNS notification.
- C
Create a CloudWatch alarm on the average CPU utilization metric of the ALB, set the threshold to 70%, and configure the alarm to trigger an Amazon SNS notification.
- D
Create a CloudWatch alarm on the average CPU utilization metric of the auto-scaling group, set the threshold to 70%, and configure the alarm to automatically scale out the group by one instance.
Show answer and explanation
Correct answer: A
Explanation
To monitor the average CPU utilization of an auto-scaling group and notify the operations team, you need to create a CloudWatch alarm on the average CPU utilization metric of the group. The alarm should include a threshold of 70%, a condition of 5 consecutive minutes, and an action to notify via Amazon SNS. Monitoring individual EC2 instances or unrelated metrics like the ALB's is not appropriate for this use case.
- A. Correct.
This is the correct option. CloudWatch allows you to monitor the average CPU utilization of an auto-scaling group. You can set a threshold of 70% and configure the alarm to send notifications via Amazon SNS when the threshold is breached.
- B. Incorrect.
This is incorrect because monitoring CPU utilization for each individual EC2 instance is inefficient and does not align with the requirement to monitor the average utilization across the auto-scaling group.
- C. Incorrect.
This is incorrect because the ALB does not directly report the CPU utilization of the instances behind it. The metric of interest here is the CPU utilization of the auto-scaling group, not the ALB.
- D. Incorrect.
This is incorrect because the question specifies notifying the operations team, not automatically scaling out the auto-scaling group. While scaling actions can be triggered by CloudWatch alarms, that is not the requirement in this scenario.