An application on Amazon EC2 instances in multiple Availability Zones pulls messages from an Amazon SQS queue. During bursts, backlog grows while average CPU on the Auto Scaling group stays low. Target tracking on average CPU does not scale out quickly enough. Which solution resolves this in the MOST reliable way?
Multiple choice — select one answer
Answer choices
Architectural breakdown & explanation
Queue depth per running instance reflects worker backlog; CPU is a weak signal for queue consumers.
Correct answer(s): d. Publish a custom Amazon CloudWatch metric: visible SQS messages divided by in-service instances. Scale the Auto Scaling group on that metric against an acceptable target.
Exam domain: Domain 3: Resilient Cloud Solutions (15% weight on the official guide).
Official documentation & study links
Practice this certification
This page is one question from the AWS Certified DevOps Engineer – Professional bank. Run a full timed practice exam with domain-weighted random questions in your browser.