SOA-C02 Question 36
Single answerYou are managing an application hosted on Amazon EC2 instances behind an Application Load Balancer (ALB). CloudWatch metrics show that the average latency for the ALB has significantly increased, and the HealthyHostCount metric for one of the target groups has dropped to zero. Upon investigation, you find that the EC2 instances in that target group are running out of memory. What should you do to remediate this issue and improve application availability?
- A
Increase the instance size of the EC2 instances in the affected target group to provide more memory.
- B
Reduce the health check interval for the ALB to detect unhealthy instances more quickly.
- C
Configure an Auto Scaling group for the target group to scale out EC2 instances based on memory utilization.
- D
Manually restart the EC2 instances in the affected target group to free up memory.
Show answer and explanation
Correct answer: C
Explanation
The increased ALB latency and zero HealthyHostCount indicate an availability issue caused by EC2 instances running out of memory. Configuring an Auto Scaling group allows the system to automatically scale out based on memory utilization, providing a scalable and automated solution to ensure the application remains available even under high load.
- A. Incorrect.
Increasing the instance size might temporarily alleviate the issue, but it does not provide a long-term scalable solution. Manual instance resizing also requires downtime.
- B. Incorrect.
Reducing the health check interval will help detect unhealthy instances faster but does not address the root cause of memory exhaustion or improve availability.
- C. Correct.
Configuring an Auto Scaling group to scale out instances based on memory utilization directly addresses the issue by dynamically adding more capacity as memory usage increases, improving availability and performance.
- D. Incorrect.
Manually restarting the instances might temporarily free up memory, but it is not an automated or scalable way to remediate the issue in the long term.