Skip to content

v5.0

No due date 100% complete

The core changes in scope for v5 are largely centered around node groups and offloading local module functionality to the terraform-aws-eks project. The major points of change slated for v5 are:

  1. Replace the EKS manged node groups, self-managed node groups, and Fargate profiles with the implementation provided by the terraform-aws-eks module. Currently, t…

The core changes in scope for v5 are largely centered around node groups and offloading local module functionality to the terraform-aws-eks project. The major points of change slated for v5 are:

  1. Replace the EKS manged node groups, self-managed node groups, and Fargate profiles with the implementation provided by the terraform-aws-eks module. Currently, the project utilizes this module for creation of the control plane and the change will be to remove the local module implementations in favor of the EKS module implementations for node groups and Fargate profiles
  2. Replace the KMS key module and policy document and leverage the terraform-aws-eks implementation. There is one addon, aws-for-fluentbit that utilizes the local KMS key module that will require an update to leverage the terraform-aws-kms module (same module used by the EKS module)
  3. The local launch template module is removed from the project. The terraform-aws-eks project covers the launch template functionality, including support for creating just launch templates that can be utilized by Karpenter for autoscaling
  4. Remove of the experimental warnings for optional attributes and ignored changes. The end goal will be to make the applied output as clean as possible for users, however we will have to evaluate on a case by case basis for addons

Out of scope for v5 are any refactoring of the kubernetes-addons, aws-eks-teams, irsa, and emr-on-eks modules other than those changes required to support the items listed above.

An upgrade guide will be provided for this breaking change and where possible, guidance will be provided to users on how to upgrade from v4.x to v5.x to avoid as much disruption as possible. However, due to the nature of Terraform as well as the underlying AWS resources, we cannot guarantee an upgrade path that is completely disruption free, but all efforts will be provided to minimize disruptions to cluster workloads.

Loading