██████╗ ██████╗ ██████╗ ███████╗    ████████╗ ██████╗ 
  ██╔════╝██╔═══██╗██╔══██╗██╔════╝    ╚══██╔══╝██╔═══██╗
  ██║     ██║   ██║██║  ██║█████╗         ██║   ██║   ██║
  ██║     ██║   ██║██║  ██║██╔══╝         ██║   ██║   ██║
  ╚██████╗╚██████╔╝██████╔╝███████╗       ██║   ╚██████╔╝
   ╚═════╝ ╚═════╝ ╚═════╝ ╚══════╝       ╚═╝    ╚═════╝ 

   ██████╗██╗      ██████╗ ██╗   ██╗██████╗ 
  ██╔════╝██║     ██╔═══██╗██║   ██║██╔══██╗
  ██║     ██║     ██║   ██║██║   ██║██║  ██║
  ██║     ██║     ██║   ██║██║   ██║██║  ██║
  ╚██████╗███████╗╚██████╔╝╚██████╔╝██████╔╝
   ╚═════╝╚══════╝ ╚═════╝  ╚═════╝ ╚═════╝ 

Building an AI-Powered Infrastructure Wiki with GitHub Copilot

TL;DR: This blog demonstrates how to use GitHub Copilot (Opus 4.5) with Azure MCP servers to automatically generate comprehensive infrastructure documentation from live Azure resources—turning natural language prompts into a full wiki in minutes.

---

📖 What Is This?

This repository showcases a practical approach to automating infrastructure documentation using generative AI. Instead of manually documenting Azure resources, we use GitHub Copilot to query live infrastructure and generate:

• Architecture diagrams
• Resource inventories
• Integration guides
• Operational runbooks
• Dependency maps

The result? A complete infrastructure wiki generated through 6 conversational prompts.

---

🎯 Who Is This For?

Developers

• Learn how AI assistants can accelerate documentation tasks
• Understand patterns for integrating GitHub Copilot into your workflow
• See practical examples of prompt engineering for infrastructure tasks

Platform Engineers

• Discover how to auto-generate documentation for Azure landing zones
• Explore patterns for maintaining living documentation
• Learn RBAC requirements for AI-assisted infrastructure queries

DevOps / SRE Teams

• Automate runbook generation from existing infrastructure
• Keep documentation synchronized with actual deployments
• Reduce documentation debt with AI assistance

---

🛠️ Prerequisites

Before you can replicate this workflow, ensure you have the following set up:

Option 1: GitHub Codespaces (Recommended)

The fastest way to get started—no local setup required:

What's included:

• Ubuntu minimal base image
• Azure CLI pre-installed
• GitHub Copilot + Azure MCP extensions
• 2 vCPU / 4GB RAM / 16GB storage (Codespaces Lite)

After launch, just authenticate:

az login --use-device-code

Option 2: Local Development

Tools Required

Tool	Purpose	macOS	Windows
VS Code	Development environment	brew install --cask visual-studio-code	winget install Microsoft.VisualStudioCode
GitHub Copilot	AI assistant (with Chat enabled)	Subscribe	Subscribe
Azure CLI	Azure authentication	brew install azure-cli	winget install Microsoft.AzureCLI
Azure MCP Extension	Copilot-to-Azure integration	Install via VS Code Extensions	Install via VS Code Extensions

Quick Install (Copy & Paste)

macOS (Homebrew):

brew install --cask visual-studio-code
brew install azure-cli

Windows (winget):

winget install Microsoft.VisualStudioCode
winget install Microsoft.AzureCLI

MCP Servers Used

This workflow leverages the Model Context Protocol (MCP) to connect GitHub Copilot with Azure services:

MCP Server	Capabilities	Used For
Azure Resource Graph	Query resources across subscriptions	Discovering and inventorying resources
Azure Resource Manager	List subscriptions, resource groups	Navigation and scoping
Azure CLI Integration	Execute az commands	Authentication and advanced queries

Azure RBAC Requirements

To run the prompts in this guide, your Azure identity needs the following permissions:

Role	Scope	Purpose
Reader	Subscription or Resource Group	Query resource metadata and configurations
Resource Graph Reader	Subscription	Execute Azure Resource Graph queries

Minimum Required: Reader role at the resource group level is sufficient for documentation generation.

Assigning RBAC (if needed)

# Assign Reader role at resource group scope
az role assignment create \
  --assignee <your-user-or-service-principal-id> \
  --role "Reader" \
  --scope /subscriptions/<subscription-id>/resourceGroups/<resource-group-name>

Authentication Setup

# Login to Azure CLI
az login

# Verify your subscription access
az account list --output table

# Set your target subscription
az account set --subscription "<subscription-name-or-id>"

---

🚀 The Prompt Journey

We generated this entire wiki using just 6 natural language prompts. See the complete prompt sequence, what GitHub Copilot did at each step, and lessons learned:

📄 View the Full Prompt Journey →

Highlights:

• Progressive discovery from subscription → resource group → detailed docs
• Iterative refinement of architecture diagrams
• Single prompt that generated 12 documentation files

---

📚 Generated Documentation

The following wiki was auto-generated by GitHub Copilot from the rg-artagent-voice-agent-dev resource group:

Section	Description
Architecture Overview	System design and component relationships
Data Flow	How data moves through the system
Resource Inventory	Complete Azure resource listing
Integration Guides	Service integration patterns
Azure AI Integration	AI Services configuration
Communication Services	ACS setup and usage
Dependencies	Runtime and service dependencies
Deployment Runbook	Deployment procedures
Troubleshooting Guide	Common issues and solutions

---

🏗️ The Documented Architecture

This wiki documents a Real-Time Audio Voice Agent infrastructure—a fictional scenario showcasing typical Azure AI workloads:

What the Architecture Does

The Real-Time Audio Voice Agent enables:

• 🎤 Real-time voice conversations with AI-powered agents
• 🔊 Speech-to-text and text-to-speech processing
• 📞 Voice calling capabilities via Azure Communication Services
• 📧 Email integration for notifications and follow-ups
• 💬 Web chat interface as an alternative interaction channel

High-Level Architecture

┌─────────────────────────────────────────────────────────────────────────────────┐
│                              👤 CLIENT LAYER                                     │
│  ┌─────────────────────────────────┐    ┌─────────────────────────────────┐     │
│  │       🌐 Web Browser            │    │       📞 Phone / PSTN           │     │
│  └─────────────────────────────────┘    └─────────────────────────────────┘     │
└─────────────────────────────────────────────────────────────────────────────────┘
                    │                                      │
                    ▼                                      ▼
┌─────────────────────────────────────────────────────────────────────────────────┐
│                     📦 AZURE CONTAINER APPS ENVIRONMENT                          │
│  ┌─────────────────────┐  ┌─────────────────────┐  ┌─────────────────────┐      │
│  │      Frontend       │  │      Backend        │  │      WebChat        │      │
│  │   (rtaudio-client)  │─▶│   (rtaudio-server)  │◀─│      (Demo UI)      │      │
│  └─────────────────────┘  └─────────────────────┘  └─────────────────────┘      │
└─────────────────────────────────────────────────────────────────────────────────┘
                                       │
          ┌────────────────────────────┼────────────────────────────┐
          ▼                            ▼                            ▼
┌───────────────────────┐  ┌───────────────────────┐  ┌───────────────────────────┐
│  🤖 AZURE AI SERVICES │  │ 📡 COMMUNICATION SVC  │  │   ⚙️ PLATFORM SERVICES    │
├───────────────────────┤  ├───────────────────────┤  ├───────────────────────────┤
│ • AI Foundry Hub      │  │ • Communication Svc   │  │ • Key Vault               │
│ • AI Project          │  │ • Email Services      │  │ • App Configuration       │
│ • Speech Services     │  │ • PSTN Integration    │  │ • Container Registry      │
└───────────────────────┘  └───────────────────────┘  └───────────────────────────┘

📖 Detailed diagrams available in docs/architecture/overview.md

---

🌐 Network Architecture

This section details the networking configuration discovered by querying the live Azure resources.

Network Topology Summary

Aspect	Configuration	Notes
VNet Integration	❌ Not configured	Container Apps Environment uses public networking
Private Endpoints	❌ None deployed	All services use public endpoints
NSGs	❌ Not applicable	No custom VNet = no NSG requirements
Subnets	❌ Not applicable	Managed by Azure Container Apps

⚠️ Development Environment: This architecture uses public networking suitable for dev/test. Production deployments should implement VNet integration and private endpoints.

Service Endpoints & Access

Container Apps (Compute Layer)

Service	FQDN	Port	External
Frontend	<frontend-app>.<environment-id>.eastus2.azurecontainerapps.io	8080	✅ Yes
Backend	<backend-app>.<environment-id>.eastus2.azurecontainerapps.io	8000	✅ Yes
WebChat	<webchat-app>.<environment-id>.eastus2.azurecontainerapps.io	3001	✅ Yes

Container Apps Environment

Property	Value
Static IP	<static-ip-address>
Default Domain	<environment-id>.eastus2.azurecontainerapps.io
Public Network Access	Enabled
VNet Configuration	None (Azure-managed networking)
Zone Redundancy	Disabled
mTLS	Disabled

Data Services Network Configuration

Service	Type	Public Access	Private Endpoints	Network Rules
Key Vault	kv-<suffix>	Enabled	None	RBAC Authorization enabled
Storage Account	st<suffix>	Enabled	None	Bypass: AzureServices, Default: Allow
AI Services	<project><suffix>aif	Enabled	None	None
Redis Enterprise	redis<suffix>	Enabled	None	TLS 1.2 minimum
Cosmos DB (Mongo)	cosmos-cluster-<suffix>	Enabled	None	None

Redis Enterprise Details

Property	Value
Host	<redis-name>.eastus2.redis.azure.net
Port	10000
SKU	MemoryOptimized_M10
Redis Version	7.4
High Availability	Enabled
Redundancy	Zone Redundant (ZR)
TLS Version	1.2+
Access Keys Auth	Disabled (Entra ID)

Cosmos DB Mongo Cluster Details

Property	Value
Connection String	mongodb+srv://<user>:<password>@<cluster-name>.mongocluster.cosmos.azure.com/
Server Version	8.0
SKU	M30
Storage	128 GB
High Availability	Disabled
Replication Role	Primary

---

💡 Key Takeaways

What We Learned

1. Natural Language Works — Conversational prompts like "what's in this resource group?" are effective for infrastructure discovery

2. Progressive Discovery — Start broad (subscription) and narrow down (resource group → specific resources)

3. Iterative Refinement — Diagrams and documentation improve through follow-up prompts

4. Context Matters — Providing LLM instruction files (.github/copilot-instructions.md) improves future AI interactions

Limitations to Consider

• AI-generated docs should be reviewed for accuracy
• Sensitive information (keys, connection strings) must be filtered
• Complex architectures may need multiple prompt iterations
• Real-time data means docs can become stale

---

🔄 Try It Yourself

Quick Start

1. Clone this repo to see the generated output
2. Review the Prompt Journey to understand the process
3. Set up prerequisites (see above)
4. Run these prompts against your own Azure infrastructure:

1. "Do we have access to Azure? List my subscriptions and resource groups."

2. "What resources are in [your-resource-group]?"

3. "Create a wiki documenting [resource-group] with architecture diagrams, 
    resource inventory, integration guides, dependencies, runbooks, 
    and LLM instruction files following best practices."

4. [Iterate on specific sections as needed]

---

📊 Results Summary

Metric	Value
Total Prompts	6
Files Generated	13
Azure Queries	6
Documentation Pages	10
Time to Generate	~15 minutes

---

🔐 Security Considerations

When using AI to document infrastructure:

• ✅ Reader-only access is sufficient—no write permissions needed
• ✅ No secrets exposed — AI queries metadata, not secret values
• ✅ Audit trail — All queries go through Azure Resource Graph
• ⚠️ Review outputs — Ensure no sensitive resource names are shared publicly

---

📁 Repository Structure

├── README.md                           # This file (blog/guide)
├── PROMPT_JOURNEY.md                   # Detailed prompt documentation
├── .github/
│   └── copilot-instructions.md         # GitHub Copilot context file
└── docs/                               # Generated wiki documentation
    ├── architecture/
    ├── resources/
    ├── integrations/
    ├── dependencies/
    └── runbooks/

---

🤝 Contributing

Found this useful? Have improvements?

• ⭐ Star this repo if you found it helpful
• 🐛 Open an issue for questions or suggestions
• 🔀 Submit a PR with improvements to the prompts or documentation

See CONTRIBUTING.md for guidelines.

---

📜 License

This project is licensed under the MIT License. Feel free to use, modify, and share.

---

✍️ Author

Kevin Evans — Code to Cloud

Built with GitHub Copilot (Claude Opus 4.5) + Azure MCP Servers — February 2026