为 Autoscaling Group(CF) 中的 EC2 实例设置 CloudWatch 警报

Question

I have an AWS::AutoScaling::AutoScalingGroup configuration that runs two instances of EC2. My question is - is it possible to attach CloudWatch alarms for both instances? For example I want to observe StatusCheckFailed_Instance metric for each EC2 in a group?

Usually you can attach alarms through the EC2 Instance ID but how to know each EC2 Instance ID in AutoScalingGroup to attach alerts? or here should be another way to attach alerts? I really can't find something useful and workable over internet.

Answer 1

Option 1)

• Create your own script that's triggered on launch/terminate events
• the scripts will each be set to trigger a lambda that would read in the instance ID and create/delete an alarm

Option 2)

• If you're not trying to use the auto-recover option (which you shouldn't need in an ASG, since the ASG will just replace the instances), then you can make 1 aggregate alarm for the ASG
• Create the alarm based on the StatusCheckFailed_Instance metric with the ASGName=<> Dimmension
• Set it to trigger if the MAX statistic value is > 1 (since that means at least 1 instance is failing, each instance will push its own datapoints to ASG versions of EC2 metrics)
• Since you only have 2 instances, you can just manually check both if it ever triggers. But for larger ASGs using the SEARCH() math expression on the CloudWatch metrics console (or a dashboard) would be a good way to look through all the ASG instances and view their metrics to see which one is failing