Open mitchtchesnitch opened 1 year ago
👋
thanks for your interest and the kind words 🙂
Am I maybe missing something or is this a point of improvement?
To be perfectly honest, we haven't looked into all the details of the implementation yet. So far we basically fixed the tool for our use-case and updated the documentation accordingly. So I couldn't tell you if there is a valid reason for the wait time after updating the last node.
If you feel like it you could do some digging, we would gladly review any PR :)
Hi!
Thank you so much for forking and maintaining this project. This is an absolute godsend for folks that want to use EKS but find that its node rolling update mechanism is too opinionated for stateful workloads or ones that take a long time to recover after pod eviction.
I've been playing around with the tool and I am making use of the BETWEEN_NODES_WAIT wait. I noticed that this wait will also be added right after the last node has been worked on and is shut down, leading to an un-necessary wait in the tail-end of the operation.
Am I maybe missing something or is this a point of improvement?
Thank you!