SOA-C02 Question 39
Single answerYou are managing a fleet of EC2 instances behind an Auto Scaling group that is serving a web application. An Amazon CloudWatch alarm has triggered, indicating that the average CPU utilization of the instances in the Auto Scaling group has exceeded 80% for the past 5 minutes. What is the most appropriate corrective action to take in this scenario?
- A
Manually add more EC2 instances to the Auto Scaling group.
- B
Verify the Auto Scaling group's scaling policies and ensure the desired capacity can be increased.
- C
Check the application logs on the instances to identify the root cause of the high CPU usage before taking any action.
- D
Terminate underperforming instances and allow the Auto Scaling group to replace them with new instances.
Show answer and explanation
Correct answer: B
Explanation
When a CloudWatch alarm indicates high CPU utilization across an Auto Scaling group, the immediate corrective action should be to ensure the Auto Scaling group has appropriate scaling policies to handle the increased load. This allows the system to automatically add more instances to distribute the load and maintain performance. Manually intervening or troubleshooting logs might be helpful later but does not resolve the immediate alarm condition effectively.
- A. Incorrect.
Manually adding instances to the Auto Scaling group is not an efficient or scalable solution, as the purpose of Auto Scaling is to handle scaling automatically based on defined policies.
- B. Correct.
Ensuring the Auto Scaling group has appropriate scaling policies and that the desired capacity can increase is the correct action, as this aligns with the goal of maintaining application performance when resource demand increases.
- C. Incorrect.
While checking application logs might help diagnose longer-term issues, it does not address the immediate problem of high CPU usage and will not mitigate the alarm condition in a timely manner.
- D. Incorrect.
Terminating underperforming instances is not a recommended approach in this situation, as it does not guarantee that the new instances will resolve the high CPU issue. Instead, scaling out is the appropriate response.