Your team creates a Docker image for a new web application and deploys it in an AWS ECS cluster. You want to automatically increase or decrease the desired count of ECS tasks in the ECS service based on the average CPU utilization. The number of ECS tasks should be adjusted between 3 and 6 and the target value of CPU utilization should be 70. How would you configure this?

Answer options:

A.Add a target tracking scaling policy in the Auto Scaling group used by the AWS ECS cluster. Make sure that the policy uses the ECSServiceAverageCPUUtilization metric.
B.In the AWS ECS service console, add two step scaling policies in the Auto Scaling tab. Set up two CloudWatch alarms to scale in and scale out based on the average CPU.
C.Set up the target tracking auto scaling in the ECS task definition. Configure the minimum, desired and maximum number of tasks based on the ECSTaskAverageCPUUtilization metric in CloudWatch.
D.Configure the auto scaling in the ECS service by adding a target tracking scaling policy based on the ECSServiceAverageCPUUtilization metric.

Correct Answer – D
Option A is incorrect because the auto scaling should be configured in the ECS service. Although you can configure a scaling policy in the Auto Scaling group, the scaling policy cannot adjust the number of ECS tasks.
Option B is incorrect because in the ECS service, you should configure a target tracking scaling policy instead of step scaling policies. The question mentioned that the value of CPU utilization should be targeted as 70 so that a target tracking scaling policy is needed.
Option C is incorrect because the target tracking scaling policy should be configured in the ECS service instead of the ECS task definition.
Option D is CORRECT because you can add a scaling policy for the ECS service. The number of ECS tasks will automatically scale based on the target value of the ECSServiceAverageCPUUtilization metric.
Reference:
https://docs.aws.amazon.com/AmazonECS/latest/developerguide/service-autoscaling-targettracking.html.