Support for azure ml v2 Kubernetes cluster attachments

[zargaryanlilit]

Is there an existing issue for this?

• [X] I have searched the existing issues

Community Note

• Please vote on this issue by adding a :thumbsup: reaction to the original issue to help the community and maintainers prioritize this request
• Please do not leave "+1" or "me too" comments, 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

Description

There is an azurerm_machine_learning_inference_cluster for the azure ml v1 AKS cluster attachments. Azure ml v2 K8s cluster attachment is different e.g. we need to provide the 'namespace'.

New or Affected Resource(s)/Data Source(s)

azurerm_machine_learning_kubernetes_cluster

Potential Terraform Configuration

resource "azurerm_machine_learning_inference_cluster" "example" {
  name                  = "example"
  location              = azurerm_resource_group.example.location
  namespace             = "default"
  kubernetes_cluster_id = azurerm_kubernetes_cluster.example.id
  description           = "This is an example cluster used with Terraform"

  machine_learning_workspace_id = azurerm_machine_learning_workspace.example.id

  tags = {
    "stage" = "example"
  }
}

References

No response

[walonso]

Hi, I'm interested in this feature, is this already released, is there a plan to have it and release it?

[alegend4u]

Currently, resource azurerm_machine_learning_inference_cluster by default considers the legacy AksCompute which only supports v1 SDK of Azure ML.

Additional argument (maybe compute_type) for resource azurerm_machine_learning_inference_cluster with possible values AksCompute and KubernetesCompute would do.

Refer official doc: KubernetesCompute vs Legacy AksCompute

[lyubomirk]

I am also interested in this feature. I have installed an AzureML extension, on which AKS compute is not supported, and currently there is no option for Kubernetes compute.