Azure SRE Agent is Microsoft’s AI-driven incident-response agent, built into Azure rather than deployed as a separate open-source project. Where What is HolmesGPT is a bring-your-own-infra CNCF Sandbox project you wire into whatever platform you already run, Azure SRE Agent is Azure-native by design — it starts investigating the moment an alert fires, without a human triggering it.
Architecture: alert-triggered agentic loop
Alert fires (PagerDuty / ServiceNow / Azure Monitor)
│
▼
SRE Agent acknowledges the alert
│
▼
Queries Azure Monitor, App Insights, Kusto, connected 3rd-party tools
│
▼
Hypothesis-driven investigation (extended reasoning for complex incidents)
│
▼
Root cause + proposed fix ──▶ executes or waits for approval, per configured autonomy
Incident Response Plans — routing and autonomy control
Response plans route an incoming incident to the right custom agent, at the right autonomy level, based on rules you define — filterable by severity, service, title, and type. This is the configuration surface: instead of one agent handling every incident the same way, different incident shapes get routed to different plans.
Run modes
| Mode | Behavior |
|---|---|
| Assisted | Investigates and proposes a fix; a human approves before anything executes |
| Autonomous | For scenarios trusted and well-tested enough to skip approval — the agent remediates directly |
Learning loop: Session Insights
Unlike a static runbook, the agent generates a Session Insight after every incident — a structured summary of what happened, what worked, and what to improve — and that record feeds forward into how future incidents of the same shape get handled. The “learns and remembers” framing is the same design goal What is HolmesGPT and other agentic-RCA tools are converging on; the difference is where that memory lives and who controls it.
Where it sits next to HolmesGPT
| Concern | Azure SRE Agent | What is HolmesGPT |
|---|---|---|
| Vendor / openness | Microsoft, Azure-native, closed | Robusta.dev, Apache 2.0, CNCF Sandbox |
| Integration surface | Azure-first: Azure Monitor, App Insights, Kusto, PagerDuty/ServiceNow/Azure Monitor incidents | 70+ toolsets across Kubernetes, VMs, cloud, databases, any infra |
| Autonomy control | Per Incident Response Plan — Assisted vs. Autonomous | ”Operator mode” — continuous background monitoring, opens fix PRs |
| Memory | Session Insights per incident | No equivalent persistent per-incident memory layer documented |
Why it’s relevant here: the per-plan, configurable autonomy model (route by severity/service, choose assisted vs. autonomous per route) is a cleaner articulation of the same human-in-the-loop-by-design tension that Build an AI SRE Assistant (Part 00) works through with its own approval-gate design — worth comparing Microsoft’s routing-based autonomy control directly against that pattern, alongside HolmesGPT’s read-only-RBAC approach to the same problem, before extending that hands-on build further.
Local graph
Linked from 3 notes
Managing, Monitoring, and Securing Azure AI Services
Operational reference for running Azure AI Services in production — activity logs vs. diagnostic logs vs. metrics, cost model and budgets, key rotation and Key Vault, Entra ID auth, network isolation (service vs. private endpoints), Responsible AI principles, CI/CD and IaC for AI resources, and container deployment.
What is Azure AI Services
Microsoft's Azure AI service catalog — account models (single- vs multi-service), Azure OpenAI's deployment-based access pattern, Azure AI Search as the RAG grounding layer, and the single-service capability catalog (Vision, Language, Speech, Document Intelligence).
Agentic AI: Projects & Engineering Mastery
A book-shaped table of contents for Agentic AI: Projects & Engineering Mastery: hands-on practitioner builds, Principal/Staff-level technical leadership, and the lookup appendices and vendor/framework reference notes for the whole series. Book 6 of the AI Systems Engineering series.
Related notes
What is Azure AI Services
Microsoft's Azure AI service catalog — account models (single- vs multi-service), Azure OpenAI's deployment-based access pattern, Azure AI Search as the RAG grounding layer, and the single-service capability catalog (Vision, Language, Speech, Document Intelligence).
Managing, Monitoring, and Securing Azure AI Services
Operational reference for running Azure AI Services in production — activity logs vs. diagnostic logs vs. metrics, cost model and budgets, key rotation and Key Vault, Entra ID auth, network isolation (service vs. private endpoints), Responsible AI principles, CI/CD and IaC for AI resources, and container deployment.
What is HolmesGPT
Robusta.dev's open-source SRE agent (CNCF Sandbox) for investigating production incidents across Kubernetes, VMs, cloud services, and databases — an agentic tool-calling loop over 70+ toolsets, not a chatbot or RAG system, with a proactive 'operator mode' that monitors and opens fix PRs without a human trigger.
What is gcx
Grafana's official CLI for managing Grafana Cloud/Enterprise/OSS resources, optimized for agentic usage — dashboards, alerts, SLOs, metrics/logs/traces/profiles queries, and every major Cloud product, plus a bundled Agent Skills set for Claude Code and other .agents-compatible harnesses.