AZ-900 is a conceptual exam, but the easiest way to retain the material is to connect each concept to a small Azure deployment. The objective is not to become an Azure engineer in one week; it is to understand the platform’s building blocks and the language used to describe them.
Understand the exam domains
| Skill area | Weight | What matters |
|---|---|---|
| Cloud concepts | 25–30% | Shared responsibility, public/private/hybrid models, consumption pricing, serverless, benefits and IaaS/PaaS/SaaS. |
| Azure architecture and services | 35–40% | Regions, availability zones, subscriptions, resource groups, compute, networking, storage, identity, security and service selection. |
| Azure management and governance | 30–35% | Cost management, tags, Azure Policy, locks, RBAC, the portal, Cloud Shell, ARM/Bicep concepts, Azure Monitor and service health. |
Seven-day plan
Shared responsibility, CapEx versus OpEx, consumption pricing, elasticity, scalability, high availability and cloud service models.
Regions, paired regions, availability zones, subscriptions, management groups, resource groups and resources.
Virtual machines, scale sets, containers, App Service, Functions, virtual networks, subnets, VPN, ExpressRoute, Load Balancer, Application Gateway and DNS.
Storage accounts, Blob, Files, queues, tables, redundancy, access tiers, Entra ID, RBAC, MFA, Conditional Access and Zero Trust concepts.
Defender for Cloud, Key Vault, Azure Policy, resource locks, tags, budgets, Cost Management and the Cloud Adoption Framework.
Portal, Cloud Shell, Azure CLI, PowerShell, ARM templates, Bicep, Azure Arc, Azure Monitor, Advisor and Service Health.
Take a timed assessment, revisit weak comparisons and practise explaining why one service is a better fit than another.
Comparisons worth memorising
| Comparison | Key distinction |
|---|---|
| IaaS vs PaaS vs SaaS | How much of the stack the customer manages versus the provider. |
| Availability zones vs region pairs | Separate datacentre zones inside a region versus paired regions used for broader resiliency. |
| RBAC vs Azure Policy | Who can perform actions versus what resource configurations are allowed. |
| Resource lock vs Policy | Protection from delete/change operations versus continuous governance and compliance evaluation. |
| Load Balancer vs Application Gateway | Layer 4 traffic distribution versus Layer 7 web routing and web application features. |
| Blob vs Azure Files | Object storage accessed through APIs/HTTP versus managed SMB or NFS file shares. |
Hands-on exercises
- Create a resource group and apply tags.
- Deploy a low-cost storage account and inspect redundancy and access-tier options.
- Open Cloud Shell and list resource groups with Azure CLI and PowerShell.
- Create a budget or explore the cost-analysis blade.
- Review Azure Policy built-in definitions and a sample initiative.
- Open Azure Service Health and compare it with Resource Health and Azure Monitor.
az group list --output table
az account show --output table
Get-AzContext
Get-AzResourceGroup | Select-Object ResourceGroupName, LocationExam strategy
- Read for the required outcome: cost, availability, management effort, security or governance.
- Fundamentals questions favour the service definition that best matches the scenario.
- Eliminate answers from the wrong service layer before comparing similar options.
- Do not over-engineer; the exam is testing foundational service selection.
Final readiness checklist
- I can draw the Azure management hierarchy.
- I can explain core compute, networking and storage services.
- I can distinguish RBAC, Policy, locks, tags and management groups.
- I completed Microsoft Learn and a timed Practice Assessment.
Microsoft references
- Official AZ-900 study guide
- Microsoft Certified: Azure Fundamentals
- Course AZ-900T00: Introduction to Cloud Infrastructure
- Azure documentation
- Microsoft Certification Practice Assessments
Review note: Certification objectives, Microsoft cloud features and portal labels change. Check the official Microsoft page for the date of your exam or production deployment.