Normal for EC2 Autoscaling Public AMIs to terminate a machine every hour?


I have some behavior happening in our newly spun up EC2 Autoscaled AMI setup that is not expected. I’ll try and explain what is going on as best I can:

  1. The Min and Max settings seem to work as the min is always correct and we never go above the max.
  2. For some reason unknown to me we always seem to have a recently terminated instance in AWS and a relatively new one I assume spun up in its place.
  3. Unfortunately, this one instance that is spun up in its place does not seem to respond to testing queues.
  4. We actually have to cause the autoscaling to kick in for a new instance, which then will pick up the queue correctly.
  5. This newly spun up instance seems to work fine, but once we slow our testing, things will get terminated and the process begins again.
  6. I’m a bit of an AWS novice so this may be an AWS specific thing that I am not familiar with.

Some questions:

  1. Is this normal behavior?
  2. If not, does anyone know of any settings we may have set incorrectly off the top of their head?
  3. Regarding the default location setting in settings.ini - Is it possible to have multiple locations in there as well? We would like to have individual autoscaling per location as they all have different needs.
  4. Is there some AWS setting that I can check that might explain this behavior?

Any help would be highly appreciated, thank you!