🚀 Enhancements
Include new AMI types AL2023_x86_64_(NVIDIA|NEURON) @m-wynn (#202)
## what- Amazon has some new AMI types
- AL2023_x86_64_NEURON
- AL2023_x86_64_NVIDIA
- These follow the same SSM parameters
/aws/service/eks/optimized-ami/1.31/amazon-linux-2023/x86_64/neuron/amazon-eks-node-al2023-x86_64-neuron-1.31-v20250103/image_id/aws/service/eks/optimized-ami/1.31/amazon-linux-2023/x86_64/nvidia/amazon-eks-node-al2023-x86_64-nvidia-1.31-v20250103/image_id/aws/service/eks/optimized-ami/1.31/amazon-linux-2023/x86_64/neuron/recommended/image_id/aws/service/eks/optimized-ami/1.31/amazon-linux-2023/x86_64/nvidia/recommended/image_id
why
We want to be able to deploy all types of AMIs into our cluster for GPU and ML purposes. These are the AL2023 equivalents of AL2_x86_64_GPU
references
https://github.com/awslabs/amazon-eks-ami/releases
https://aws.amazon.com/blogs/containers/amazon-eks-optimized-amazon-linux-2023-accelerated-amis-now-available/
🐛 Bug Fixes
Bugfixes @Nuru (#205)
## what- Fix use of
var.kubernetes_version - Configure Terraform to create new node group before destroying old one when switching
var.create_before_destroyfromfalsetotrue - Revert code to be compatible with Terraform v1.3
- Update test dependencies
- Better tolerance of inputs from disabled EKS cluster module
why
- Incorrectly treated list as scalar
- Better user experience to implement create-before-destroy behavior when selecting that behavior
- Version spec says module is compatible with Terraform 1.3
- Keep current
- Reduce special handling needed by client modules when setting
enabled=false