Quota - implemented AZD quota check, removed model array for embedding and gpt only, removed scripts

Author: Seth

azure.yaml

@@ -14,15 +14,4 @@ hooks:
       shell: sh
       run: chmod u+r+x ./scripts/set_conns_env_vars.sh; ./scripts/set_conns_env_vars.sh
       interactive: true
-      continueOnError: false
-  preprovision:
-    posix:
-      shell: sh
-      run: chmod u+r+x ./scripts/validate_model_deployment_quotas.sh; chmod u+r+x ./scripts/validate_model_quota.sh; ./scripts/validate_model_deployment_quotas.sh --subscription $AZURE_SUBSCRIPTION_ID --location $AZURE_LOCATION --models-parameter "aiModelDeployments"
-      interactive: false
-      continueOnError: false
-    windows:
-      shell: pwsh
-      run: ./scripts/validate_model_deployment_quotas.ps1 -Subscription $env:AZURE_SUBSCRIPTION_ID -Location $env:AZURE_LOCATION -ModelsParameter "aiModelDeployments"
-      interactive: false
       continueOnError: false

infra/main.bicep

@@ -5,11 +5,27 @@ targetScope = 'resourceGroup'
 @description('The name of the environment/application. Use alphanumeric characters only.')
 param name string
 
-@description('Specifies the location for all the Azure resources. Defaults to the location of the resource group.')
+@metadata({ azd: { type: 'location' } })
+@description('Specifies the location for all the Azure resources.')
 param location string
 
-@description('Optional. Specifies the OpenAI deployments to create.')
-param aiModelDeployments deploymentsType[] = []
+@description('Specifies the AI embedding model to use for the AI Foundry deployment. This is the model used for text embeddings in AI Foundry. NOTE: Any adjustments to this parameter\'s values must also be made on the aiDeploymentsLocation metadata in the main.bicep file.') 
+param aiEmbeddingModelDeployment modelDeploymentType
+
+@description('Specifies the AI chat model to use for the AI Foundry deployment. This is the model used for chat interactions in AI Foundry. NOTE: Any adjustments to this parameter\'s values must also be made on the aiDeploymentsLocation metadata in the main.bicep file.')
+param aiGPTModelDeployment modelDeploymentType
+
+@metadata({
+  azd: {
+    type: 'location'
+    usageName: [
+      'OpenAI.GlobalStandard.gpt-4o,150'
+      'OpenAI.GlobalStandard.text-embedding-3-small,100'
+    ]
+  }
+})
+@description('Required. Location for AI Foundry deployment. This is the location where the AI Foundry resources will be deployed.')
+param aiDeploymentsLocation string
 
 @description('Specifies whether creating an Azure Container Registry.')
 param acrEnabled bool = false
@@ -83,12 +99,6 @@ param documentIntelligenceEnabled bool = false
 param networkAcls object ={
   defaultAction: 'Deny'
   bypass: 'AzureServices' // ✅ Allows trusted Microsoft services
-  // virtualNetworkRules: [
-  //   {
-  //     id: networkIsolation ? network.outputs.vmSubnetName : ''
-  //     ignoreMissingVnetServiceEndpoint: true
-  //   }
-  // ]
 }
 
 @description('Name of the first project')
@@ -216,13 +226,26 @@ module cognitiveServices 'modules/cognitive-services/main.bicep' = {
   params: {
     name: name
     resourceToken: resourceToken
-    location: location
+    location: aiDeploymentsLocation
     networkIsolation: networkIsolation
     networkAcls: networkAcls
     virtualNetworkResourceId: networkIsolation ? network.outputs.virtualNetworkId : ''
     virtualNetworkSubnetResourceId: networkIsolation ? network.outputs.vmSubnetId : ''
     logAnalyticsWorkspaceResourceId: logAnalyticsWorkspace.outputs.resourceId
-    aiModelDeployments: aiModelDeployments
+    aiModelDeployments: [
+      for model in [aiEmbeddingModelDeployment, aiGPTModelDeployment]: {
+        name: empty(model.?name) ? model.modelName : model.?name
+        model: {
+          name: model.modelName
+          format: 'OpenAI'
+          version: model.version
+        }
+        sku: {
+          name: 'GlobalStandard'
+          capacity: model.capacity
+        }
+      }
+    ]
     userObjectId: userObjectId
     contentSafetyEnabled: contentSafetyEnabled
     visionEnabled: visionEnabled
@@ -357,7 +380,7 @@ module sqlServer 'modules/sqlServer.bicep' = if (sqlServerEnabled) {
   }
 }
 
-import { sqlDatabaseType, databasePropertyType, deploymentsType } from 'modules/customTypes.bicep'
+import { sqlDatabaseType, databasePropertyType, modelDeploymentType } from 'modules/customTypes.bicep'
 
 
 output AZURE_KEY_VAULT_NAME string = keyvault.outputs.name

infra/main.parameters.json

@@ -50,33 +50,19 @@
     "bingGroundingEnabled": {
       "value": "${AZURE_BING_GROUNDING_ENABLED}"
     },
-    "aiModelDeployments": {
-      "value": [
-        {
-          "name": "textembed",
-          "model": {
-            "name": "text-embedding-ada-002",
-            "format": "OpenAI",
-            "version": "2"
-          },
-          "sku": {
-            "name": "Standard",
-            "capacity": 10
-          }
-        },
-        {
-          "name": "gpt",
-          "model": {
-            "name": "gpt-4o",
-            "version": "2024-05-13",
-            "format": "OpenAI"
-          },
-          "sku": {
-            "name": "GlobalStandard",
-            "capacity": 10
-          }
-        }
-      ]
+    "aiEmbeddingModelDeployment": {
+      "value": {
+        "modelName": "text-embedding-3-small",
+        "version": "1",
+        "capacity": 100
+      }
+    },
+    "aiGPTModelDeployment": {
+      "value": {
+        "modelName": "gpt-4o",
+        "version": "2024-05-13",
+        "capacity": 150
+      }
     }
   }
 }

infra/modules/customTypes.bicep

@@ -253,6 +253,22 @@ type longTermBackupRetentionPolicyType = {
   yearlyRetention: string?
 }
 
+@export()
+@description('The AI model deployment type for Cognitive Services account.')
+type modelDeploymentType = {
+  @description('Optional. The name of the Cognitive Services account deployment model. The modelName will be used by default if not specified.')
+  name: string?
+
+  @description('Required. The format of the Cognitive Services account deployment model.')
+  modelName: string
+
+  @description('Required. The version of the Cognitive Services account deployment model.')
+  version: string
+
+  @description('Required. The capacity of the resource model definition representing SKU.')
+  capacity: int
+}
+
 @export()
 type deploymentsType = {
   @description('Optional. Specify the name of cognitive service account deployment.')