Open devopsmash opened 2 years ago
Any updates on this maybe? We've been experiencing some issues with Cluster autoscaler and AZRebalance
Same issue for me.
I can't help you on coding it, but it I was able to do it with a null_resource command using
aws autoscaling suspend-processes --auto-scaling-group-name ${each.value} --scaling-processes AZRebalance
Auto scaling group names can be accessed through terraform-provider-aws module (I used a data source however for simplicity).
Unfortunately, there's no support for passing down the desired suspended_processes
through to the underlying ASG in the EKS API. Given that the underlying ASG isn't managed by Terraform, I think that @eclivus' suggestion is the best option right now. If anyone interested in this feature has a support contract with AWS it might be worth you trying to engage with your TAM to see if the feature can be put on the EKS backlog. There's a related roadmap item https://github.com/aws/containers-roadmap/issues/1453 (with a similar null_resource
workaround) which it might be worth folks tracking and/or providing input to.
Community Note
Description
while the resource
aws_autoscaling_group
hassuspended_processes
like:The resource
aws_eks_node_group
doesn't supportsuspended_processes
It seems that the Managed Autoscale group has this option
Can you add this support for the resource
aws_eks_node_group
?New or Affected Resource(s)