02/06/2023, 3:34 PM
In an autoscaling cluster, sometimes nodes time out during registration. This results in it being stuck in it being there forever, even if the instance is removed by the ASG in AWS. Is there any way for rancher to be able to set some check to clean this up after a certain time. I've seen up to hundreds of nodes at times that never go away and have no instances backing them. Since they don't register, they aren't listed in kubectl, so it has to be done internally in rancher somehow.