Microsoft’s Azure SRE Agent: AI-Powered Reliability Engineering

Posted on September 14, 2025

Microsoft’s Azure SRE Agent is a newly introduced AI-powered tool designed to enhance site reliability engineering (SRE) practices within Azure environments.

🌐 Overview

Launched in May 2025, the Azure SRE Agent helps teams sustain production cloud environments by automating incident response, diagnostics, and infrastructure monitoring. It leverages large language models (LLMs) to analyze logs, metrics, and configurations, enabling faster root cause analysis and issue mitigation.

⚙️ Core Capabilities

  • Incident Response Automation: Reduces toil by handling repetitive tasks like log analysis and alert triage.
  • Root Cause Analysis (RCA): Uses AI to identify anomalies and suggest remediation steps.
  • Continuous Learning: Builds contextual understanding of your Azure resources to provide intelligent insights.
  • Natural Language Interface: You can ask questions like:
    • “What changed to my app in the last day?”
    • “Can you give me the overall AKS cluster usage?”

🧠 Operational Modes

The agent operates in three distinct modes:

  1. Read-only: Observes and reports without making changes.
  2. Review: Suggests actions but requires user approval to execute.
  3. Autonomous: Executes actions independently, ideal for mature environments.

🔐 Security & Setup

  • Requires Role-Based Access Control (RBAC) permissions.
  • Currently available only in the Sweden Central region during preview.
  • Setup involves creating a dedicated resource group and configuring access via Azure CLI.

🚀 Integration & Benefits

  • Seamlessly integrates with Azure observability tools and GitHub Copilot.
  • Improves uptime, reduces operational costs, and frees engineers to focus on strategic tasks.

To set up the Azure SRE Agent, follow these steps:

✅ Prerequisites

Before creating the agent, ensure the following:

  1. Permissions:
    • Your account must have Microsoft.Authorization/roleAssignments/write permissions.
    • You should be a Role-Based Access Control (RBAC) Administrator or User Access Administrator.
  2. Subscription Setup:
    • Set your Azure CLI session to the correct subscription ID:
    • az account set --subscription "<SUBSCRIPTION_ID>"
  3. Namespace Registration:
    • Register the required namespace:
    • az provider register --namespace "Microsoft.App"
  4. Region Restriction:
    • The agent is currently only available in the Sweden Central region.
    • Ensure you have permissions to create resources in this region.

🛠️ Creating the Agent

  1. Access the Portal:
    • Use the link provided in your onboarding email to open the Azure SRE Agent portal.
  2. Create the Agent:
    • Click Create.
    • Fill in the required fields in the Create Agent window.
    • Create a new resource group specifically for the agent (separate from your application’s resource group).
  3. Associate Resource Groups:
    • Click Choose resource groups.
    • Select the resource groups you want the agent to monitor.
    • Click Save.
  4. Finalize Creation:
    • Click Create to deploy the agent.

⚙️ Choose a Run Mode

You can select one of three modes:

  • Read-only: Observes and reports, no changes made.
  • Review: Suggests actions, requires your approval.
  • Autonomous: Executes actions independently.

More Information : Azure SRE Agent documentation