Skip to main content
  1. Public practice questions
  2. AZ-104 Practice Exams | Azure Administrator Associate
  3. AZ-104 Case Study Set

Question 1 — AZ-104 Case Study Set

Paideffort Enterprise AI Platform Infrastructure Modernization
 

Business Context

Paideffort operates a professional certification-learning platform used by customers in several regions. The engineering organization is replacing manually provisioned Azure resources with infrastructure as code. The first workload being migrated is an AI-assisted learning service that generates explanations, retrieves course information, invokes internal APIs, and delegates selected long-running operations to Azure Functions.

Business requirements

Development and production must be deployed from version-controlled infrastructure definitions. The same deployment artifacts must support different Azure regions and environment names. Infrastructure deployments must be repeatable without creating duplicate resources. Authentication for the AI application must use Microsoft Entra ID instead of application-managed API keys. Production deployments must run through CI/CD, while engineers must retain a supported command-line method for local validation and deployment.

Technical Architecture

The application uses a Microsoft Foundry project endpoint for agent and project operations. Embedding requests are intended to use the Azure OpenAI endpoint through the OpenAI SDK. DefaultAzureCredential is used for Microsoft Entra authentication. An OpenAPI tool exposes Paideffort's course-catalog API, a FunctionTool handles deterministic application functions, and an Azure Functions integration handles queue-backed processing that can outlive an interactive request.

Application integration
from azure.identity import DefaultAzureCredential, get_bearer_token_provider
from azure.ai.projects import AIProjectClient
from openai import OpenAI

credential = DefaultAzureCredential()

project_client = AIProjectClient(
    endpoint=FOUNDRY_PROJECT_ENDPOINT,
    credential=credential
)

token_provider = get_bearer_token_provider(
    credential,
    "https://ai.azure.com/.default"
)

openai_client = OpenAI(
    base_url=FOUNDRY_PROJECT_ENDPOINT,
    api_key=token_provider
)

embedding = openai_client.embeddings.create(
    model=EMBEDDING_DEPLOYMENT,
    input=query_text
)

Infrastructure Implementation

The platform team created an ARM template and committed it beside the application source. Deployment-time values are intended to vary between development and production. Storage is required for application artifacts and Azure Functions processing. The current template contains the following storage resource:

{
  "$schema": "https://schema.management.azure.com/schemas/2019-04-01/deploymentTemplate.json#",
  "contentVersion": "2.1.0.0",
  "parameters": {
    "environmentName": {
      "type": "string",
      "allowedValues": [ "dev", "prod" ]
    },
    "location": {
      "type": "string"
    }
  },
  "variables": {
    "storageName": "[concat('paideffort', parameters('environmentName'), uniqueString(resourceGroup().id))]"
  },
  "resources": [
    {
      "type": "Microsoft.Storage/storageAccounts",
      "apiVersion": "2025-01-01",
      "name": "[variables('storageName')]",
      "location": "westus",
      "sku": {
        "name": "[parameters('storageSku')]"
      },
      "kind": "StorageV2",
      "properties": {
        "supportsHttpsTrafficOnly": true
      }
    }
  ],
  "outputs": {
    "storageAccountName": {
      "type": "string",
      "value": "[variables('storageName')]"
    }
  }
}

The development deployment is requested for East US, while production will initially run in West Europe. The release team expects the region supplied to the deployment to determine where regional resources are created.

Deployment Pipeline

The current GitHub Actions workflow deploys the template and then provisions several additional services individually:

- name: Deploy platform
  run: |
    az group deployment create \
      --resource-group $RESOURCE_GROUP \
      --template-file ./infra/main.json

    az storage account create \
      --resource-group $RESOURCE_GROUP \
      --name $ARTIFACT_STORAGE \
      --location $LOCATION

    az functionapp create \
      --resource-group $RESOURCE_GROUP \
      --name $FUNCTION_APP \
      --storage-account $ARTIFACT_STORAGE
Operational objective

Paideffort wants deployment history to provide a consistent view of infrastructure state and deployment outputs. Engineers also want Azure Resource Manager to coordinate resource creation rather than maintaining an expanding sequence of service-specific provisioning commands.

Future Requirements

During the next release cycle, Paideffort expects separate platform, AI, data, and tool-integration components to be reused across multiple workloads. The architecture must support higher regional scale, stronger identity controls, additional agent tools, centralized deployment governance, and improved observability. The infrastructure team wants large deployments divided into maintainable components while preserving coordinated deployment through a primary infrastructure definition.