AWS Certified DevOps Engineer – Professional — Question 197

A company has an application that runs on 12 Amazon EC2 instances. The instances run in an Amazon EC2 Auto Scaling group across three Availability Zones.
On a typical day each EC2 instance has 30% CPU utilization during business hours and 10% CPU utilization after business hours. The CPU utilization increases suddenly in the first few minutes of business hours each day. Other increases in CPU utilization are gradual. A DevOps engineer needs to optimize costs while maintaining or improving the application's reliability.
Which solution meets these requirements?

Answer options

• A. Configure a target tracking scaling policy that is based on the Auto Scaling group's average CPU utilization, and set a target of 75%. Create a scheduled action for the Auto Scaling group to adjust the desired capacity to six instances just before business hours begin.
• B. Configure the Auto Scaling group with two scheduled actions for Amazon EC2 Auto Scaling. Configure one action to start nine EC2 instances at the start of business hours. Configure the other action to stop nine instances at the end of business hours.
• C. Change to an AWS Application Auto Scaling group. Configure a target tracking scaling policy that is based on the Auto Scaling group's average CPU utilization, and set a target of 75%. Create a scheduled action for the Auto Scaling group to adjust the minimum number of instances to three instances at the end of business hours and to reset the number to six instances before business hours begin.
• D. Change to an AWS Application Auto Scaling group. Configure a target tracking scaling policy that is based on the Auto Scaling group's average CPU utilization, and set a target of 75%. Create a scheduled action to terminate nine instances each evening at the end of business hours.

Correct answer: A

Explanation

Option A is correct because it employs a target tracking scaling policy that reacts to CPU utilization while also scheduling a reduction in capacity to optimize costs during off-peak hours. Options B and C involve excessive instance management without effectively leveraging the target tracking policy's responsiveness, and Option D does not consider the need for capacity during business hours, potentially degrading application reliability.