aws_eks_node_group accepts instance_types, while EKS only allows one type

[TBeijen]

Community Note

• Please vote on this issue by adding a 👍 reaction to the original issue to help the community and maintainers prioritize this request
• Please do not leave "+1" or other comments that do not add relevant new information or questions, they generate extra noise for issue followers and do not help prioritize the request
• If you are interested in working on this issue or have submitted a pull request, please leave a comment

Terraform Version

Terraform: 0.12.16
Provider aws: 2.43.0

Affected Resource(s)

• aws_eks_node_group

Terraform Configuration Files

resource "aws_eks_node_group" "workers" {
  # ...<truncated>...
  instance_types  = ["m5a.large", "m5.large"]

Debug Output

Error: instance_types: attribute supports 1 item maximum, config has 2 declared

  on .terraform/modules/eks_cluster_apps_1.eks_cluster/node_groups.tf line 69, in resource "aws_eks_node_group" "workers":
  69: resource "aws_eks_node_group" "workers" {

Expected Behavior

No error when following resource documentation

Actual Behavior

Error as displayed above.

Steps to Reproduce

Trying to create a eks_node_group resource specifying more than 1 instance type (1 works fine)

Important Factoids

This seems to be a combination of:

• Resource docs providing misguiding information
• EKS being limited to one type, as can be seen when using CLI.

Suggested changes:

• Update docs and examples
• Consider renaming var to instance_type. This would need to be aligned with EKS roadmap. If at some point EKS plans to support multiple instance types, it's better to keep this var name.

[bflad]

Hi @TBeijen 👋 Thank you for submitting this.

Our preference in the Terraform AWS Provider is to follow the API naming and structures provided by AWS. This is to prevent difficulties caused by introducing our own abstractions (for maintainers and operators switching between tooling such as the AWS CLI or CloudFormation) and for general future-proofing of Terraform resources against upstream API changes that otherwise may cause operator burden if our abstractions are incorrect and the original implementation needs to be deprecated.

In this case, the EKS CreateNodegroup API denotes the field as the plural instanceTypes array of strings, which matches the Terraform implementation.

I have submitted #11457 to update our documentation to mention the current API restriction.

[ghost]

This has been released in version 2.45.0 of the Terraform AWS provider. Please see the Terraform documentation on provider versioning or reach out if you need any assistance upgrading.

For further feature requests or bug reports with this functionality, please create a new GitHub issue following the template for triage. Thanks!

[ghost]

I'm going to lock this issue because it has been closed for 30 days ⏳. This helps our maintainers find and focus on the active issues.

If you feel this issue should be reopened, we encourage creating a new issue linking back to this one for added context. Thanks!