coultn
commented
5 years ago @adlemich is this question related to scaling the cluster? Or is it about scaling the service (i.e. number of tasks in the service)? You can actually define separate cooldown timeouts for scale in vs. scale out for both cluster and service scaling, see here for documentation:
https://docs.aws.amazon.com/cli/latest/reference/autoscaling-plans/create-scaling-plan.html
Furthermore, the behavior should already be what you are requesting - the cooldown period only blocks scaling actions in the same direction as the previous scaling action. That means, for example, that a scale out can happen immediately after a scale in, regardless of the cooldown period.
Can you describe your configuration in more detail so we can better understand what the gap is in the existing features?
Tell us about your request Scale out cool down should be different than scale in cool down and the values must not be the same. Also the cool down should only apply to scaling into the same direction. At least scaling out should happen immediately even when a scale down cool down period is still active.
Which service(s) is this request for? Fargate, ECS
Tell us about the problem you're trying to solve. What are you trying to do, and why is it hard? Currently we have the problem that a cluster which is scaling in and is in cooldown can not scale out immediately when request traffic is increasing at that time. This can lead to a dead period that is as long as the cool down.
Are you currently working around this issue? Not scaling in basically, which is expensive to operate
Additional context none
Attachments none