Question

A media company runs a batch image-processing service on Amazon ECS with AWS Fargate. Jobs arrive in an Amazon SQS queue, and worker tasks pull messages to process them. During the day, queue depth is minimal, but overnight batch uploads cause tens of thousands of messages to accumulate, and tasks fall far behind. The team wants the number of Fargate tasks to automatically match the backlog so jobs complete faster during spikes, while scaling in to reduce cost when the queue empties. Which approach best meets these requirements?

Accepted Answer

Configure ECS Service Auto Scaling with a target tracking policy based on a custom CloudWatch metric derived from SQS ApproximateNumberOfMessages per task (backlog per task).

Answer

Set a fixed desired count on the ECS service large enough to handle the overnight peak so tasks are always available.

Answer

Configure ECS Service Auto Scaling with a target tracking policy based on the average CPU utilization of the Fargate tasks.

Answer

Migrate the workers to a larger Fargate task size (vertical scaling) so each task processes more messages concurrently.