Amazon Web Services DOP-C02 Exam With Confidence Using Practice Dumps

Option B is the correct choice and has the least operational overhead because ECR lifecycle policies are the native, managed feature built specifically to automate image cleanup:

ECR lifecycle policies can expire images based on image age (e.g., images older than X days) and can also enforce retaining a minimum number of images (count-based retention), which directly matches the requirement “retain at least a specified number of images.”

Lifecycle policies work for both tagged and untagged images using rule selection and tagging status filters, so the repository stays clean without custom automation.

Why the other options are not appropriate / higher overhead:

A is invalid: ECR is not S3; S3 lifecycle policies do not manage ECR images .

C and D introduce custom automation (Lambda or SSM scripting), scheduling, permissions, error handling, retries, logging, and ongoing maintenance—higher overhead than a managed lifecycle policy.

Comprehensive and Detailed Explanation From Exact Extract of DevOps Engineer documents only:

During an Amazon ECS service deployment, if tasks fail the Elastic Load Balancing target group health checks, Amazon ECS considers those tasks unhealthy and stops and replaces them. This behavior can cause the deployment to remain incomplete because healthy replacement tasks never become available.

Also, if an essential container in the task exits, the entire task stops. Since the question states that the new tasks enter a stopped state soon after launch, an essential container failure is a direct and highly likely cause.

Why the other options are incorrect:

B. The IAM role used by the DevOps engineer to update the ECS service does not need RunTask permission for the tasks to remain running after launch. A permission issue here would more likely affect the update action itself rather than cause launched tasks to stop shortly afterward.

C. Although CloudWatch Logs configuration problems can affect logging, the missing log group is not the most likely direct reason for ECS tasks to repeatedly stop in this deployment scenario.

D. The task role provides permissions to the application code running inside the container. It is not the role primarily responsible for launching the task. Task startup is typically related to the task execution role, ECS service behavior, container health, and load balancer checks.