Course – AIOps & Agentic Operations Engineering
AIOps Engineering — Building Intelligent Operations at Scale
24 Hrs (5 Days / 5 Hrs or 6 Days / 4 Hrs) Advanced Classroom / Live Virtual 75% Hands-On
Duration
24 Hrs / 5-6 Days
Level
Advanced
Format
Classroom / Live Virtual
Stack
Azure Foundry · Agent Framework · Azure Monitor
Training Methodology
Learning by Doing
Every session in this program is built around hands-on execution, not passive slides — you leave having built and deployed working AIOps agents, not just watched a demo.
01
Explore
Hands-on labs from Day 1 — real Azure Monitor and Foundry environments, not slide decks.
02
Experiment
Work against simulated IT environments that mirror production observability and incident scenarios.
03
Engage
Build detection, RCA, self-healing and FinOps agents across every module, not toy examples.
04
Apply
Capstone builds a working end-to-end AIOps prototype tied to your team’s actual stack.
Who This Is For
What You’ll Be Able to Do
- Architect full-stack observability pipelines on Azure Monitor and design SRE-aligned SLIs, SLOs and error budgets
- Build ML-driven signal processing pipelines that correlate, group and de-noise high-volume alert streams
- Deploy topology-aware RCA agents on Azure Foundry that reason across service dependency graphs
- Build closed-loop self-healing agents with Microsoft Agent Framework, including human-in-the-loop escalation and rollback guardrails
- Secure agentic AIOps systems against prompt injection with least-privilege access, audit logging and red-teaming
- Integrate FinOps cost observability and build agents that trigger automated scale-down on cost anomalies
- Build RAG pipelines on Azure AI Search that ground RCA agents in operational runbook knowledge
- Apply MLOps practices — model registry, drift detection and automated retraining — to operationalize AIOps models
Prerequisites
- Completion of foundational Agentic AI training on the Microsoft technology stack, with working knowledge of agent design, development and deployment
- Familiarity with IT Operations processes — incident, change and service management — and basic Azure cloud computing concepts
- Hands-on experience with Azure services, Azure Foundry and Microsoft Agent Framework, plus basic Python proficiency and understanding of REST APIs and event-driven architectures
Curriculum
Module 1
AIOps Architecture & Observability Foundation
- AIOps vs traditional IT Ops — architectural shift
- Full-stack observability — logs, metrics and traces
- Azure Monitor as the observability backbone
- Telemetry data schema, normalization and instrumentation strategies
- Observability maturity model
Lab: Configure Azure Monitor to collect logs, metrics and traces from a simulated IT environment.
Module 2
SRE-Driven Reliability Design
- SRE principles and their role in AIOps
- Defining SLIs, SLOs and SLAs
- Error budget policies and burn rate alerts
- Reliability as a design constraint in AIOps
- Integrating SRE metrics into operational dashboards
Module 3
Intelligent Signal Processing
- Alert fatigue and the noise problem in IT Ops
- Event correlation techniques — rule-based and ML-driven
- ML-based noise reduction approaches
- Alert grouping and deduplication strategies
- Threshold-based vs. anomaly-based alerting
- Building signal processing pipelines on Azure
Lab: Build an event correlation pipeline that filters and groups high-volume Azure Monitor alerts into prioritized actionable signals.
Module 4
Anomaly Detection & Topology-Aware RCA
- ML model types for anomaly detection in IT Ops
- Operationalizing pre-built anomaly detection models on Azure
- Topology mapping and service dependency graphs
- Root cause analysis frameworks
- Agentic AI for automated RCA reasoning
- Integrating RCA agents with Azure Foundry
Lab: Deploy a pre-built anomaly detection model and configure an Azure Foundry RCA agent to reason across a service topology map.
Module 5
Predictive Analytics & Real-Time Telemetry Pipelines
- Forecasting models for operational failure prediction
- Real-time vs. batch telemetry processing
- Azure Event Hubs and Stream Analytics for telemetry ingestion
- Building scalable telemetry pipelines on Azure Foundry
- Operationalizing forecasting models in production
- Predictive alerting and capacity planning
Lab: Deploy a real-time telemetry ingestion pipeline with predictive alerting for capacity threshold breaches.
Module 6
Closed-Loop Self-Healing Systems
- Closed-loop automation architecture
- Detection → diagnosis → remediation design pattern
- Building self-healing agents with Microsoft Agent Framework
- Agent orchestration in multi-step remediation workflows
- Human-in-the-loop escalation patterns
- Safety guardrails and rollback mechanisms
Lab: Build a self-healing agent on Azure Foundry that detects a simulated incident, diagnoses root cause, and executes automated remediation.
Module 7
AI Agent Security & Prompt Injection Defense
- Prompt injection attacks on LLM-based agents
- Agent identity & access management — least-privilege design
- Audit logging of autonomous agent actions
- Secrets management for agents calling external APIs
- Threat modeling for agentic AIOps systems
- Security testing and red-teaming AI agents
Lab: Configure Azure Key Vault secrets management and audit logging for a self-healing agent, then simulate a prompt injection attack and validate defensive controls.
Module 8
Cost Observability & FinOps Integration
- Cost as a first-class operational metric
- Azure Cost Management API integration with Azure Monitor
- Cost anomaly detection — budget alerts and spend spikes
- Agent-triggered scale-down actions tied to cost policies
- FinOps dashboards alongside reliability dashboards
- FinOps governance and showback/chargeback models
Lab: Integrate Azure Cost Management API with Azure Monitor and build an agent that detects a cost spike and triggers an automated scale-down action.
Module 9
RAG for Runbook Knowledge
- Vector indexing of operational runbooks in Azure AI Search
- Grounding RCA agents with retrieved runbook context
- Post-mortem ingestion pipelines for agent learning
- Hybrid search — keyword and semantic for incident context
- RAG pipeline architecture on Azure Foundry
- Evaluating RAG quality — relevance, groundedness and faithfulness
Lab: Build a RAG pipeline that indexes operational runbooks in Azure AI Search and grounds an RCA agent with retrieved context during a simulated incident diagnosis.
Module 10
MLOps for AIOps Operationalization
- MLOps principles in the AIOps context
- Model registry, versioning and governance on Azure
- Model drift detection and monitoring
- CI/CD pipelines for model deployment
- Agentic AI for automated model monitoring and response
- Continuous model evaluation and retraining triggers
Lab: Configure a model monitoring pipeline where an Azure Foundry agent detects drift and triggers an automated retraining workflow.
Module 11
Capstone — Working AIOps Prototype
- Capstone problem statement and scope definition
- End-to-end system architecture design
- Integration of observability, agents, RAG and MLOps components
- Prototype build, testing and validation
- Demonstration and peer review
Delivery Details
- Delivered as classroom or live virtual instructor-led — scheduled around your team, 5 hrs/day over 5 days or 4 hrs/day over 6 days
- 75% hands-on labs across Azure Foundry, Microsoft Agent Framework and Azure Monitor
- Labs are illustrative and may vary by trainer approach, participant profile and Azure subscription access