Microsoft.MachineLearningServices/workspaces/outboundRules returns MissingApiVersionParameter

[emad0082]

Brief description of the problem

When using Microsoft.MachineLearningServices/workspaces/outboundRules@2023-10-01 as documented here to create outbound rules in the workspace. Terraform throws the following error, indicating that the api_version parameter is not set. However the documentation provides the version in the type parameter.
The request itself actually has it specified which version is being used, so i dont understand how this error happens in the first place.

When applying the terraform code, an error with the message: RESPONSE 400: 400 Bad Request
ERROR CODE: ValidationError happens. - {"error":{"code":"MissingApiVersionParameter","message":"The api-version query parameter (?api-version=) is required for all requests."}}

Terraform code:

resource "azapi_resource" "outbound_rules" {
  type = "Microsoft.MachineLearningServices/workspaces/outboundRules@2023-10-01" 
  name = "CognitiveServiceOutboundRule"
  parent_id = azurerm_machine_learning_workspace.default.id
  locks = [azurerm_machine_learning_workspace.default.id]
  body = jsonencode({
  "properties": {
    "category": "UserDefined",
    "status": "Active",
    "type": "PrivateEndpoint",
    "destination": {
      "serviceResourceId": "${azurerm_private_endpoint.openai_ple.id}",
      "sparkEnabled": false,
      "sparkStatus": "Inactive",
      "subresourceTarget": "account"
      }
    }
  })
}

Full stack trace of the error:

Error: Failed to create/update resource

  with azapi_resource.outbound_rules,
  on workspace.tf line 100, in resource "azapi_resource" "outbound_rules":
 100: resource "azapi_resource" "outbound_rules" {

creating/updating Resource: (ResourceId
"/subscriptions/xxxxx/resourceGroups/xxxxx/providers/Microsoft.MachineLearningServices/workspaces/AIP-SB-Core-mlw/outboundRules/CognitiveServiceOutboundRule"
/ Api Version "20[23](https://github.com/xxxx/actions/runs/xxxxx#step:17:24)-10-01"): PUT


{"error":{"code":"MissingApiVersionParameter","message":"The api-version query parameter (?api-version=) is required for all requests."}}
--------------------------------------------------------------------------------
RESPONSE 400: 400 Bad Request
ERROR CODE: ValidationError
--------------------------------------------------------------------------------
{
  "error": {
    "code": "ValidationError",
    "severity": null,
    "message": "Unsupported Target resource Microsoft.Network/privateEndpoints in Rule: \"CognitiveServiceOutboundRule\". Check documentation for supported types:https://aka.ms/AMLSupportedPEsInManagedNetwork",
    "messageFormat": null,
    "messageParameters": null,
    "referenceCode": null,
    "detailsUri": null,
    "target": null,
    "details": [],
    "innerError": null,
    "debugInfo": null,
    "additionalInfo": null
  },
  "correlation": {
    "operation": "xxxx",
    "request": "4717f16cd1228543"
  },
  "environment": "westeurope",
  "location": "westeurope",
  "time": "2024-05-08T08:55:20.4,
  "componentName": "account-rp",
  "statusCode": 400
}
--------------------------------------------------------------------------------

How to reproduce

Implement a private endpoint, a machine learning workspace in your terraform config.
Specify type to this version: type = "Microsoft.MachineLearningServices/workspaces/outboundRules@2023-10-01"

Run the plan.

Apply, and it should fail.

Current workarounds/fixes

As of so far i have not been able to add private endpoint outbound rules to the resource. Any help would be greatly appreciated

[emad0082]

How do i add a label to this?

[ms-henglu]

Hi @emad0082 ,

Thank you for taking time to report this issue and apologize for late response.

There's a similar issue related to ML workspace outbound rules, and I shared some examples in the comment: #453 (comment)

Please check if it could help in your case.

[emad0082]

Thank you for responding, i will be attempting to use your example fixes and return once tested.

[emad0082]

Alright @ms-henglu

I have now tested the proposed solutions. And i can only get the FQDN rules to work. The privat endpoint outbound rules still returns the same error. This is after i have tested it with version 1.12.1 of the provider and the azapi_resource_action solution.