AWS: How is CPU utilization calculated for auto scaling group

Question

I created an Auto Scaling group, with an Application Load Balancer attached, with a policy that when CPU utilization rate is greater than 75% add one more instance. It seems working.

When I have four instances running, how is that CPU Utilization rate calculated? Average of the four instances? Minimum among the four?

Answer 1

You created an Auto Scaling Policy and that policy executes when the alarm you had configured is triggered.

You can see Average selected in the dropdown list. Various option that can be selected are Average, Minimum, Maximum, Sum and SampleCount.

You can also refer to Scaling Based on Metrics.