Server Fault is a question and answer site for system and network administrators. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Asked this over at the AWS forums with no response, so I thought I'd give it a go here.

If I have two auto-scaling priorities: 25% increase 50% increase

And I have two separate alarms on CPU utilization for them: 50% CPU Utilization = 25% increase 75% CPU Utilization = 50% increase

If CPU jumps to 80%, it will trigger both alarms, but will it fire the 50% increase policy or the 25% increase policy? If there are a way to ensure it would fire the 50% increase?

There doesn't seem to be any way to prioritise alarms or policies in this way that I can see.

share|improve this question
What's the URL of your post on the AWS forum? – Eric Hammond Oct 20 '11 at 23:02

Both policies will trigger if you don't have DefaultCooldown set long enough in your config to prevent rapid events.

You can use the API call DescribeAutoScalingGroups to check what this value is set to.

Example: If this value is set to 300 (it's in seconds) that means the group is set to only scale every 5 minutes. If the 50% CPU alarm is triggered and then shortly after the 75% CPU alarm is triggered the system will wait 5 minutes before allowing the second to go through if the alarm is still present.

If the usage jump is from 10% to 80% immediately the 75% CPU alarm should trigger.

share|improve this answer
So essentially you are saying they will both trigger, and both will scale if DefaultCooldown is low (zero is possible?) That is at least easy to test. I was particularly interested whether it was determined that the 75% CPU alarm would trigger first and scale first, or is the order "undefined". – Bazman Jan 2 '13 at 10:41

From the Auto Scaling Docs (last paragraph):

... Auto Scaling chooses the policy that has the greatest impact on the Auto Scaling group. For example, suppose that the policy for CPU utilization instructs Auto Scaling to launch 1 instance, while the policy for the SQS queue prompts Auto Scaling to launch 2 instances. If the scale out criteria for both policies are met at the same time, Auto Scaling gives precedence to the SQS queue policy, because it has the greatest impact on the Auto Scaling group. This results in Auto Scaling launching two instances into the group. This precedence applies even when the policies use different criteria for scaling out. For instance, if one policy instructs Auto Scaling to launch 3 instances, and another instructs Auto Scaling to increase capacity by 25 percent, Auto Scaling give precedence to whichever policy has the greatest impact on the group at that time.

share|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.