Skip to content

AWS Update Support Policy and can't change from Standard to Extend

0

I have deployed EKS Cluster running with Kubernetes version v1.31, which the version ended on November 26th. My team think it was supposed to change from standard policy to extend policy like in AWS Docs for further usage without upgrading because some issues with upgrading EKS cluster with some all the components in EKS cluster like Karpenter and KEDA cause errors

But my question right now are:

Enter image description here

  1. First Image: I want to change to Extend Support for further usage before mitigating errors and still continuing processing workloads. But we can see I can't change, and the error shows "The Kubernetes version 1.31" has reached the end of standard support and can no longer be opted-in to extended support. ". What should I do in order to change? Or does EKS change automatically to Extended Support after AWS ends support for that version?

Enter image description here

  1. Second Image: In the "Support period" status, we can see " Pending automatic upgrade to version 1.32" => Does it mean AWS will automatic update the EKS cluster for me? If does, HOW LONG until does the automatic update take action (priority) ? Does AWS guarantee my eks clsuter updated without errors.

  2. I am struggling with upgrade EKS Cluster from 1.31 with KEDA and Karpenter isn't compatible for Kubernetes version 1.32 so upgrade EKS cluster up cause the error. So does anyone know the way or guide to upgrade the EKS cluster as well as all components in EKS like KEDA and Karpenter without any downtime.

asked 10 months ago374 views

1 Answer
0

Based on the information provided, I'll address your questions about EKS support policies and upgrades:

  1. Regarding changing to Extended Support: You cannot change from Standard to Extended support after a Kubernetes version has reached the end of standard support. The policy can only be set while the cluster is running on a Kubernetes version that's still in standard support. Since Kubernetes 1.31 has already reached the end of standard support, you can no longer opt-in to extended support for this version. The support policy must be set before the version enters extended support.

By default, new and existing clusters have extended support enabled, unless specified otherwise. If your cluster was set to Standard support before 1.31 reached end of standard support, it will remain in that state and be scheduled for automatic upgrade.

  1. Regarding "Pending automatic upgrade to version 1.32": Yes, this means AWS will automatically upgrade your EKS cluster because it's set to Standard support. When a cluster is set to Standard support, it will be automatically upgraded at the end of the standard support period. AWS doesn't provide specific timing for when the automatic upgrade will occur after the end of standard support, but the status indicates it's pending.

AWS doesn't explicitly guarantee that the automatic upgrade will occur without errors, especially if you have components like KEDA and Karpenter that might not be compatible with the newer version.

  1. Regarding upgrading with KEDA and Karpenter: Since you're experiencing compatibility issues with KEDA and Karpenter on Kubernetes 1.32, you have a few options:
  • Create a new cluster with version 1.32 or newer, test your components there, and migrate workloads once compatibility issues are resolved
  • If you need to maintain your current cluster, you'll need to update your KEDA and Karpenter versions to ones that are compatible with Kubernetes 1.32 before the automatic upgrade occurs
  • Consider reaching out to AWS Support for assistance with your specific configuration challenges

For future clusters, remember to set the upgrade policy to Extended if you want to remain on specific Kubernetes versions longer (though this comes with higher costs).
Sources
View current cluster upgrade policy - Amazon EKS
UpgradePolicyRequest - Amazon EKS
Understand the Kubernetes version lifecycle on EKS - Amazon EKS
UpgradePolicyResponse - Amazon EKS

answered 10 months ago

You are not logged in. Log in to post an answer.

A good answer clearly answers the question and provides constructive feedback and encourages professional growth in the question asker.