AVMPostDeploymentGuide.md

AVM Post Deployment Guide

📋 Note: This guide is specifically for post-deployment steps after using the AVM template. For complete deployment from scratch, see the main Deployment Guide.

---

This document provides guidance on post-deployment steps after deploying the Document Knowledge Mining Solution Accelerator from the AVM (Azure Verified Modules) repository.

Overview

After deploying the infrastructure using AVM, you'll need to complete the application layer setup, which includes:

• Configuring Kubernetes infrastructure
• Building and deploying container images
• Setting up certificates and ingress controllers
• Configuring application settings

Prerequisites

Before starting the post-deployment process, ensure you have the following:

Required Software

1. PowerShell (v7.0+ recommended, v5.1+ minimum) - Available for Windows, macOS, and Linux

2. Azure CLI (v2.50+) - Command-line tool for managing Azure resources

3. kubectl - Kubernetes command-line tool
   Install using Azure CLI:

   az aks install-cli

4. aks-preview extension - Azure CLI extension for AKS preview features

   az extension add --name aks-preview

5. Helm (v3.0+) - Package manager for Kubernetes

6. Docker Desktop - Container platform for building and publishing images

   ⚠️ Important: Ensure Docker Desktop is running before executing the deployment script

Azure Requirements

7. Azure Access - One of the following roles on the subscription:

   • Owner
   • User Access Administrator

8. Microsoft.Compute Registration - Verify this resource provider is registered:

   1. Navigate to Azure Portal
   2. Go to your Azure subscription
   3. Select Settings → Resource Providers
   4. Search for Microsoft.Compute and ensure it's Registered
   

9. Deployed Infrastructure - A successful Document Knowledge Mining Accelerator deployment from the AVM repository

Post-Deployment Steps

Step 1: Clone the Repository

First, clone this repository to access the post-deployment scripts:

git clone https://github.com/microsoft/Document-Knowledge-Mining-Solution-Accelerator.git

cd Document-Knowledge-Mining-Solution-Accelerator

Step 2: Run the Post-Deployment Script

The post-deployment process is automated through a single PowerShell script that completes the following tasks in approximately 20-30 minutes:

What the Script Does:

1. Configure Kubernetes Infrastructure - Set up AKS cluster settings and networking
2. Update Configuration Files - Generate FQDN, container image paths, and certificate settings
3. Deploy Kubernetes Components - Install Cert Manager, Ingress Controller, and application pods
4. Build and Push Container Images - Compile and publish application containers to Azure Container Registry
5. Provide Access Information - Display the final application URL and next steps

Execute the Script:

1. Navigate to the deployment directory:

   cd .\Deployment\

2. Choose the appropriate command based on your deployment method:

   If you deployed using custom templates, ARM/Bicep deployments, or az deployment group commands:

   .\resourcedeployment.ps1 -ResourceGroupName "<your-resource-group-name>"

   If you deployed using azd up command:

   .\resourcedeployment.ps1

   Note: Replace <your-resource-group-name> with the actual name of the resource group containing your deployed Azure resources.

   💡 Tip: Since this guide is for AVM deployments, you'll most likely use the first command with the -ResourceGroupName parameter.

3. If you encounter execution policy issues, use this alternative command:

   powershell.exe -ExecutionPolicy Bypass -File ".\resourcedeployment.ps1" -ResourceGroupName "<your-resource-group-name>"

Step 3: Provide Required Information

During script execution, you'll be prompted for:

Email Address for SSL Certificates

• Purpose: Used for issuing SSL certificates through Let's Encrypt
• Requirement: Must be a valid email address

Azure Authentication

• You'll be prompted to authenticate with Azure
• Select the appropriate Azure account
• Ensure you have the required permissions on the target subscription

Step 4: Deployment Completion

Upon successful completion, you'll see a success message with important information:

🎉 Congratulations! Your deployment is complete.

Important: Save the application URL displayed in the console - you'll need it for data upload and accessing the application.

Next Steps

1. Configure Azure OpenAI Rate Limits

⚠️ Critical: The default deployment creates models with minimal TPM (Tokens Per Minute) limits. You must increase these limits for proper performance.

Recommended TPM Thresholds

Model Name	Recommended TPM	Minimum TPM
gpt-4.1-mini	100K TPM	10K TPM
text-embedding-3-large	200K TPM	50K TPM

⚠️ Warning: Insufficient quota will cause failures during document upload and processing. Ensure adequate capacity before proceeding.

How to Update TPM Limits:

1. Navigate to Azure AI Foundry:

   • Go to Azure AI Foundry
   • Select your project
   • Go to Deployments section

2. Update Each Model:

   
   • Select each deployed model
   • Click Edit deployment
   • Increase the Tokens per Minute Rate Limit
   

3. Apply Changes: Save the configuration for each model

2. Upload and Process Sample Documents

After configuring the TPM limits, upload sample documents:

1. Navigate to the deployment directory (if not already there):

   cd .\Deployment\

2. Execute the upload script with the URL from the deployment output:

   .\uploadfiles.ps1 -EndpointUrl https://kmgs<your-dns-name>.<datacenter>.cloudapp.azure.com

3. Alternative execution (if you encounter policy issues):

   powershell.exe -ExecutionPolicy Bypass -File ".\uploadfiles.ps1" -EndpointUrl https://kmgs<your-dns-name>.<datacenter>.cloudapp.azure.com

3. Access Your Application

Once document upload is complete, you can:

• Access the web application at the URL provided in the deployment output
• Start exploring the knowledge mining capabilities
• Upload additional documents through the web interface