Provisioning of Azure Cognitive Service resources with Terraform

The Azure Cognitive services can be provisioned with Terraform building blocks using the “azurerm~>2.0” plugin.

The Azure AI cognitive services which support currently provisioning via Terraform are listed

    - Azure Computer Vision
    - Azure Face API
    - Azure Text Analytics
    - Azure Text Translator
    - Azure QnA Maker
    - Azure Personalizer
    - Azure Custom Vision
    - Azure LUIS API
    - Azure Immersive reader
    - Azure Form recognizer
    - Azure Content Moderator
    - Azure Anomaly Detector 
    - Azure Speech                  

**The Azure Cognitive Services in preview as of writing are not included

The terraform configuration starts with generic terraform “resource” declaration block followed by the type of resource & terraform resource name. Then it’s required to provide the respective Azure resource attributes like

• Name of the Azure cognitive service instance
• Dependency [with Azure Resource Group]
• Location of the Azure Resource [can be same location of Azure resource group or different]
• Kind [ which type of azure cognitive resource like Computer Vision, Face, ContentModerator, TextAnalytics etc.]
• SKU_name [pricing tier of the resource]
• Tags [if required]

The sample terraform code for provisioning the Azure Computer Vision API is as followed

resource “azurerm_cognitive_account” “computer_vision” {

depends_on = [azurerm_resource_group.example]

name = “${var.computer_vision_instance_name}”

location = azurerm_resource_group.example.location

resource_group_name = azurerm_resource_group.example.name

kind = “ComputerVision”

sku_name = “F0”

tags = {

“Environment” = “Dev”

“CostCenter” = “ML/AI”

}

}

The repository link of the Azure cognitive services APIs provisioning through Terraform configurations is available in the following link on Github

imcuteani/awesome-terraform-azcognitive

# Happy terraforming!