https://github.com/microsoftcloudessentials-learninghub/azurearcrecommendations-ai-agent
This repository provides an example of how to create a AI Agent using a basic architecture (without network isolation). It is designed for quick demos and should be customized as needed to meet specific use cases or requirements.
https://github.com/microsoftcloudessentials-learninghub/azurearcrecommendations-ai-agent
azure-agent azure-arc disaster-recovery
Last synced: 2 months ago
JSON representation
This repository provides an example of how to create a AI Agent using a basic architecture (without network isolation). It is designed for quick demos and should be customized as needed to meet specific use cases or requirements.
- Host: GitHub
- URL: https://github.com/microsoftcloudessentials-learninghub/azurearcrecommendations-ai-agent
- Owner: MicrosoftCloudEssentials-LearningHub
- License: mit
- Created: 2025-08-04T13:55:30.000Z (2 months ago)
- Default Branch: main
- Last Pushed: 2025-08-04T19:18:45.000Z (2 months ago)
- Last Synced: 2025-08-04T19:31:26.660Z (2 months ago)
- Topics: azure-agent, azure-arc, disaster-recovery
- Homepage:
- Size: 13.7 KB
- Stars: 0
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Demo: Automating Recommendations from Azure Arc with an AI Agent (full-code approach)
Costa Rica
[](https://github.com)
[](https://github.com/)
[brown9804](https://github.com/brown9804)
Last updated: 2025-07-30
-----------------------------
`Arc API → Function App → AI Foundry → Logic Apps → Monitoring`
> - Function App is the central orchestrator for ingestion and enrichment.
> - AI Foundry provides decision intelligence.
> - Logic Apps executes or escalates actions.
> - Monitoring ensures observability and compliance.
> [!IMPORTANT]
> Disclaimer: This repository contains example of how to automate the recommendations from Azure Arc by introducing an AI-driven agent that not only ingests and processes recommendations but also:
> - Classify & priority-rank each recommendation
> - Summarize actionable next steps in natural language
> - Decide `auto-execute` vs `human-review` paths
> This is `just a guide`. It is not an official solution. For official guidance, support, or more detailed information. Please refer [RAG with Zero-Trust – Architecture Reference to Microsoft's official documentation](https://github.com/Azure/GPT-RAG) or contact Microsoft directly: [Microsoft Sales and Support](https://support.microsoft.com/contactus?ContactUsExperienceEntryPointAssetId=S.HP.SMC-HOME)
| **Category** | **Components**| **Purpose** |
|--------------------------------|-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|-------------------------------------------------------------------------------------------------------------------|
| **Core Components** | - **Azure Arc API**
- **Resource Group**
- **Subscription**| - Source of recommendations (DR, security, performance, compliance) for on-prem and hybrid assets.
- Groups all resources under a single RG and subscription scope. |
| **Data Engineering Pipeline** | - **Function App**
*Every Function App requires a General-Purpose v2 Storage Account for triggers, state, and logging.*
- **App Service Plan** (Consumption/Premium SKU)
*The App Service Plan can be serverless (Consumption) or a dedicated tier (Premium/Dedicated).*
- **Storage Account** (General Purpose v2 for Functions runtime) | Hosts and scales your ingestion/enrichment logic; fetches recommendations and sends them to AI for processing. |
| **AI Layer** | **AI Foundry**| Classifies severity, summarizes actions, prioritizes recommendations, and suggests auto-execute vs manual review. |
| **Automation & Orchestration** | **Logic Apps**| Executes safe actions (DR failover, patching, SQL fixes) or sends Teams/Email approvals for high-risk items. |
| **Monitoring & Governance** | - **Azure Monitor + Log Analytics Workspace**
- **Power BI**| Tracks pipeline health, AI decisions, execution outcomes; visualizes trends, compliance, and automation SLAs. |
## Overview
Workflow details (Click to expand)
1. Azure Arc API (Source)
- Acts as the entry point for all recommendations (DR, security, performance, compliance).
- Provides raw JSON data about advisories from on-prem and hybrid resources.
2. Function App (with App Service Plan + Storage Account): Ingest and process recommendations.
- Periodically calls Azure Arc API to fetch recommendations.
- Stores raw data temporarily in the Storage Account.
- Sends the data to the AI layer for enrichment.
3. AI Foundry: Adds intelligence to the pipeline.
- Receives raw recommendations from the Function App.
- Uses LLM models to:
- Classify severity (High/Medium/Low).
- Summarize recommendations in plain language.
- Suggest whether to auto-execute or require manual review.
- Returns enriched recommendations back to the Function App for storage and orchestration.
4. Logic Apps: Orchestrates actions based on AI decisions.
- Reads enriched recommendations.
- If `autoExecute = true`, triggers automation tasks (e.g., DR failover, patching, SQL index creation).
- If `manualReview = true`, sends Teams or email notifications for approval.
5. Monitoring & Governance:
- **Azure Monitor + Log Analytics Workspace**:
- Collects telemetry from Function App, Logic Apps, and AI calls.
- Tracks pipeline health, execution outcomes, and AI decision logs.
- **Power BI**: Connects to Log Analytics or SQL data to visualize.
- Number of recommendations processed.
- Auto-executed vs manual approvals.
- SLA compliance and risk reduction trends.
Refresh Date: 2025-08-05