A transaction processing company runs weekly scripted batch jobs on Amazon EC2 instances in an Auto Scaling group. The number of transactions varies, but baseline CPU utilization during each run is at least 60%. The company must provision capacity 30 minutes before the jobs start. Engineers currently modify the Auto Scaling group manually and the company lacks resources to analyze capacity trends. Which solution will automate modifying the Auto Scaling group’s desired capacity with the LEAST operational overhead?

Predictive scaling is the most suitable solution because it uses machine learning to forecast future traffic and proactively scales capacity. This directly addresses the requirement to provision capacity 30 minutes before the jobs start and the lack of resources to analyze capacity trends. By setting the scaling metric to CPU utilization and a target value of 60%, it ensures optimal resource allocation. The pre-launch setting ensures instances are ready in advance. Dynamic scaling policies react to current metrics, which would be too late for the pre-provisioning requirement. Scheduled scaling requires manual configuration of capacity, which the company lacks resources to analyze. Using EventBridge and Lambda adds unnecessary operational overhead and complexity compared to a native Auto Scaling policy.