Open
Description
Describe the bug
EKS: EKS uses AL2_x86_64 AmiType in AWS::EKS::Nodegroup instead of AL2023_x86_64_STANDARD
Regression Issue
- Select this option if this issue appears to be a regression.
Last Known Working CDK Version
No response
Expected Behavior
I would expect the latest AMI to be used instead of the old one. Once a cluster is deployed, a message is shown in AWS console: "New AMI release version is available for this node group."
Terraform module also suggests that "Starting on 1.30, AL2023 is the default AMI type for EKS managed node groups"
https://registry.terraform.io/modules/terraform-aws-modules/eks/aws/20.29.0
Current Behavior
The following template is generated:
"EksClusterNodegroupDefaultCapacityE6D932C8": {
"Type": "AWS::EKS::Nodegroup",
"Properties": {
"AmiType": "AL2_x86_64",
...
Reproduction Steps
const cluster = new eks.Cluster(this, 'EksGpuCluster', {
vpc: vpc,
vpcSubnets: [{ subnetType: ec2.SubnetType.PUBLIC }],
version: eks.KubernetesVersion.V1_31,
defaultCapacity: 2,
defaultCapacityInstance: new ec2.InstanceType('m5.xlarge'),
kubectlLayer: new KubectlV31Layer(this, 'KubectlLayer'),
outputConfigCommand: true,
outputClusterName: true,
authenticationMode: eks.AuthenticationMode.API_AND_CONFIG_MAP,
});
Possible Solution
No response
Additional Information/Context
No response
CDK CLI Version
2.167.1 (build d681b12)
Framework Version
No response
Node.js Version
20
OS
Mac
Language
TypeScript
Language Version
No response
Other information
No response