I'm trying to configure AWS Auto Scaling Group scale-down policy.

I've already configured the ASG to increase the number of instances if average CPU increases above a certain threshold and it works as expected but I'm having trouble configuring the opposite policy.

I'd like to scale-in incase average CPU goes below a certain threshold.

I've created some test policies, this is the scale-up policy that I've configured:

scale-up policy

This is the scale-in policy: scale-in policy

If I use the CloudWatch policy which is configured with a threshold already, then how should I configure the "take the action" part? also, it always checks for ">=" and I can't change it to "<="...

1 Answer 1


The check is a "<=" (I assume you overlooked the -infinity instead of +infinity) - so the value you put in is the upper bound in that case. ("Scale in as soon as CPUUtilization is between your value and -infinity")

Just to mention that here: you could also use the "target tracking policy" - with this you can say that auto-scaling should scale in a way that your target value is matched - that includes scale-out AND scale-in out-of-the-box.

Hope this helps!

  • +1 on target tracking if your new to the service. It works very well for most standard usecases and is extremely simple to setup. Adds or removes dynamic numbers of instances based on how extreme the metric value is at that time
    – Shahad
    Mar 18, 2020 at 1:35

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .