awslabs / amazon-eks-serverless-drainer Goto Github PK

I checked the config of sam, found that cloudwatch listen to the events EC2 Spot Instance Interruption Warning and EC2 Instance-terminate Lifecycle Action, but after decreasing the desired capacity of my eks autoscaling group, looks like there is no such two events fired

as I listen to all event of ec2, and forward all event to my email box, I did receive some other events but EC2 Spot Instance Interruption Warning and EC2 Instance-terminate Lifecycle Action in my email

make sam-logs-tail got error botocore.errorfactory.ResourceNotFoundException: An error occurred (ResourceNotFoundException) when calling the FilterLogEvents operation: The specified log group does not exist.

any idea why there is no event EC2 Spot Instance Interruption Warning and EC2 Instance-terminate Lifecycle Action fired?

Code parses the cluster name from ec2 instance tag according to this line - https://github.com/awslabs/amazon-eks-serverless-drainer/blob/master/main.sh#L37-L38

But when setting kubeconfig, it's using cluster_name variable which is different from the parsed cluster name (input_cluster_name) - https://github.com/awslabs/amazon-eks-serverless-drainer/blob/master/main.sh#L41

By default, when you deploy with a click of a button, it doesn't set the cluster_name variable at all. And now I am wondering how is this expected to work?
Also note: The lambda function itself doesn't handle errors well and makes everything look like it has worked

• to write a CDK sample to provision the refarch
• scope down the AWS Lambda role