We're currently using a mix of AWS managed Node Groups (for the system pods) and Karpenter node pools for workbench workloads. We need to switch to using one method of management to simplify and consolidate, and also offer additional features that node groups do not provide, such as automatic expiration of nodes to force AMI updates.
Related to https://positpbc.atlassian.net/browse/PTDC-181
Once this ticket is done, we need to upgrade affected customers to rotate nodes on at least a monthly basis.