I've also installed Longhorn and checked the box for "kubernetes cluster autoscaler enabled" and I'm not seeing the same behavior in that ASG, those nodes are being left alone. I have not installed the cluster autoscaler yet but I plan on it after I get this resolved.