Optimal Approach for Triggering Recreation of all EKS Managed Nodegroups
As I plan a recreation for all my EKS nodes, my preference if possible of course is to leverage the capabilities of managed nodegroups to streamline the process. Considering the various methods available, I'm seeking insights on an especially elegant and efficient approach to trigger this recreation.
PS: The goal is to restart all deployments in my EKS production cluster, we have hundreds of deployments and more than 30 nodes.
What strategies or best practices have you found effective in orchestrating a seamless rolling restart using EKS-managed nodegroups?
- 최신
- 최다 투표
- 가장 많은 댓글
You can leverage the EKS dataplane upgrade mechanism to create a gradual recreation of the node groups. This mechanism has a "Node Group update configuration." that allows you to define the following constraints on the recreation process
Number – Select and specify the number of nodes in your node group that can be updated in parallel. These nodes will be unavailable during update.
Percentage – Select and specify the percentage of nodes in your node group that can be updated in parallel. These nodes will be unavailable during update. This is useful if you have many nodes in your node group.
You can also use pod disruption budgets to maintain pod availability. see https://docs.aws.amazon.com/eks/latest/userguide/update-managed-node-group.html https://docs.aws.amazon.com/eks/latest/userguide/managed-node-update-behavior.html https://aws.github.io/aws-eks-best-practices/upgrades/#configure-poddisruptionbudgets-and-topologyspreadconstraints-to-ensure-availability-of-your-workloads-while-the-data-plane-is-upgraded
