[eks] [request]: Support logical grouping of MNGs for updates #1865
Labels
EKS Managed Nodes
EKS Managed Nodes
EKS
Amazon Elastic Kubernetes Service
Proposed
Community submitted issue
Community Note
Tell us about your request
I'd like to be able to logically group multiple MNGs together to be treated as a single unit when it comes to updates; this would match the guidance to create a MNG per AZ when running stateful workloads. Alternatively if MNGs could create an ASG per AZ behind the scenes and manage them as one this would also work. For a managed service I shouldn't have to think about this.
Which service(s) is this request for?
EKS
Tell us about the problem you're trying to solve. What are you trying to do, and why is it hard?
When I update my MNGs an instance per group is terminated which can causes a race condition where nodes can't terminate due to PDBs and pods being spread across multiple nodes. This is especially noticeable when running a node per AZ with stateful workloads. When this occurs not only does it disrupt the services but it locks some of them up due to the PVs becoming orphaned.
Are you currently working around this issue?
I can't use MNGs while this behaviour isn't supported.
Additional context
See #1866
Attachments
n/a
The text was updated successfully, but these errors were encountered: