Resource azurerm_cognitive_account_rai_policy requires severity_threshold where does not have an effect

[davidjbng]

Is there an existing issue for this?

• I have searched the existing issues

Community Note

While transitioning from azapi_resource to azurerm_cognitive_account_rai_policy I noticed the severity_threshold attribute in is required although it does not seem to have any effect for the following content filter types:

Jailbreak
Indirect Attack
Protected Material Text
Protected Material Code

Azure OpenAI Service

Looking at the Azure OpenAi settings it does not seem possible to specify values for these filters:

• 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 comments along the lines of "+1", "me too" or "any updates", 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 and review the contribution guide to help.

Terraform Version

1.10.5

AzureRM Provider Version

4.17.0

Affected Resource(s)/Data Source(s)

azurerm_cognitive_account_rai_policy

Terraform Configuration Files

resource "azurerm_cognitive_account_rai_policy" "content_filter_policy" {
  content_filter {
    name               = "Jailbreak" # or "Indirect Attack", "Protected Material Text", "Protected Material Code"
    filter_enabled     = true
    block_enabled      = true
    # severity_threshold = "High" <- value required in provider but not used in azure
    source             = "Prompt"
  }
}

Debug Output/Panic Output

2025-01-31T16:22:07.374+0100 [DEBUG] provider: plugin exited
2025-01-31T16:22:07.390+0100 [ERROR] vertex "module.azure_openai.azurerm_cognitive_account_rai_policy.content_filter_policy" error: Missing required argument
2025-01-31T16:22:07.398+0100 [DEBUG] provider.stdio: received EOF, stopping recv loop: err="rpc error: code = Unavailable desc = error reading from server: EOF"
2025-01-31T16:22:07.398+0100 [INFO]  provider: plugin process exited: plugin=.terraform/providers/registry.terraform.io/hashicorp/kubernetes/2.35.1/darwin_arm64/terraform-provider-kubernetes_v2.35.1_x5 id=98966
2025-01-31T16:22:07.398+0100 [DEBUG] provider: plugin exited
╷
│ Error: Missing required argument
│ 
│   on ../modules/azure-openai/azure-openai.tf line 94, in resource "azurerm_cognitive_account_rai_policy" "content_filter_policy":
│   94:   content_filter {
│ 
│ The argument "severity_threshold" is required, but no definition was found.
╵
2025-01-31T16:22:07.448+0100 [DEBUG] provider.stdio: received EOF, stopping recv loop: err="rpc error: code = Unavailable desc = error reading from server: EOF"
2025-01-31T16:22:07.449+0100 [INFO]  provider: plugin process exited: plugin=.terraform/providers/registry.terraform.io/hashicorp/azurerm/4.17.0/darwin_arm64/terraform-provider-azurerm_v4.17.0_x5 id=98968
2025-01-31T16:22:07.449+0100 [DEBUG] provider: plugin exited

Expected Behaviour

severity_threshold is not required for content_filter with names:

• Jailbreak
• Indirect Attack
• Protected Material Text
• Protected Material Code

Actual Behaviour

severity_threshold is required for all content_filter names

Steps to Reproduce

No response

Important Factoids

No response

References

• Support for content filters in Azure OpenAI #22822
• New Resource: azurerm_cognitive_account_rai_policy #28013