Let's say something like this is being returned by the kubectl CLI.

~]$ kubectl version
error: exec plugin: invalid apiVersion "client.authentication.k8s.io/v1alpha1"

Or by Terraform.

~]$ terraform apply -auto-approve
Error: Kubernetes cluster unreachable: exec plugin: invalid apiVersion "client.authentication.k8s.io/v1alpha1"

The apiVersion should be listed in your kube config file.

- name: arn:aws:eks:us-east-1:123456789012:cluster/my-cluster
  user:
    exec:
      apiVersion: client.authentication.k8s.io/v1alpha1
      args:
      - --region
      - us-east-1
      - eks
      - get-token
      - --cluster-name
      - my-cluster
      command: aws
      env:
      - name: AWS_PROFILE
        value: default

Some versions of the AWS CLI should set the apiVersion to be v1alpha1 and some versions of the AWS CLI should set the apiVersion to be v1beta1.

For example, if you download and install version 2.1.1 of the AWS CLI and then use the aws eks update-kubeconfig command to create or update your kube config file, apiVersion in your kube config file should be v1alpha1.

~]$ /path/to/aws-cli-version-2.1.1 eks update-kubeconfig --name my-cluster
Added new context arn:aws:eks:us-east-1:123456789012:cluster/my-cluster to /home/john.doe/.kube/config

~]$ grep apiVersion /home/john.doe/.kube/config
      apiVersion: client.authentication.k8s.io/v1alpha1

On the other hand, if you download and install version 2.13.30 of the AWS CLI and then use the aws eks update-kubeconfig command to create or update your kube config file, apiVersion in your kube config file should be v1beta1.

~]$ /path/to/aws-cli-version-2.13.30 eks update-kubeconfig --name my-cluster
Added new context arn:aws:eks:us-east-1:123456789012:cluster/my-cluster to /home/john.doe/.kube/config

~]$ grep apiVersion /home/john.doe/.kube/config
      apiVersion: client.authentication.k8s.io/v1beta1

If you are getting this error with the kubectl CLI, you can also try downloading a different version of the kubectl CLI.

curl --location --remote-name https://storage.googleapis.com/kubernetes-release/release/v1.23.6/bin/linux/amd64/kubectl

If you have two or more different kube config files, you can try pointing to a specific kube config file.

~]$ kubectl version --kubeconfig /path/to/.kube/config

And then you might want to see if the kubectl api-versions command returns output.

~]$ kubectl api-versions | grep -i authentication
authentication.k8s.io/v1

And if the kubectl api-resources command returns output.

~]$ kubectl api-resources | grep -i authentication
NAME                              SHORTNAMES   APIVERSION                               NAMESPACED   KIND
tokenreviews                                   authentication.k8s.io/v1                 false        TokenReview

If you are getting this error with Terraform, you can try a specific version of the hashicorp/aws and hashicorp/kubernetes and gavinbunney/kubectl providers.

~]$ cat required_providers.tf 
terraform {
  required_providers {

    aws = {
      source  = "hashicorp/aws"
      version = "4.67.0"
    }

    kubernetes = {
      source  = "hashicorp/kubernetes"
      version = "2.0.0"
    }

    kubectl = {
      source = "gavinbunney/kubectl"
      version = "1.9.4"
    }

  }
}