Question

A media company runs a video transcoding fleet on an EC2 Auto Scaling group behind an Application Load Balancer. During major live events, incoming jobs can spike from 20% to 90% CPU utilization within a couple of minutes, and the current target-tracking policy scales too gradually, causing a growing job backlog. The CloudOps engineer needs a scaling approach that adds capacity in proportion to how severe the breach is—launching more instances for larger spikes and fewer for minor ones. Which scaling configuration best meets this requirement?

Accepted Answer

Replace the target-tracking policy with a step scaling policy that defines multiple step adjustments based on the magnitude of the CloudWatch alarm breach.

Answer

Keep the target-tracking policy but lower the target CPU value to 40% so scaling starts earlier.

Answer

Configure a scheduled scaling action that increases desired capacity at fixed times each day.

Answer

Enable a simple scaling policy that adds a fixed number of instances whenever CPU exceeds 70%, with a 300-second cooldown.