AZ-900 for data engineers: the cloud foundation under Databricks
AZ-900 is the entry-level Azure exam. For anyone running Databricks on Azure, it organizes the foundation under every workspace: subscriptions, identity, networking and cost.
by Gabriel Henrique · · 2 min read · Ler em português
I'm Gabriel Henrique, a Data Engineer, and I work with Azure and Databricks every day. In March 2026 I earned the Microsoft Certified: Azure Fundamentals (AZ-900).
It is an introductory exam, and I still recommend it to people already working with data in the cloud. The reason is simple: many pipeline questions have their answer in a basic cloud concept.
The three parts of the exam
1. Cloud concepts
IaaS, PaaS and SaaS, public, private and hybrid cloud, the consumption model and the shared responsibility model: what Microsoft is responsible for and what the customer is responsible for.
2. Azure architecture and services
Regions, availability zones, resource groups, subscriptions and management groups, plus compute, networking, storage and identity services.
3. Management and governance
Microsoft Entra ID, role-based access control (RBAC), Azure Policy, resource locks and cost management.
Where this shows up in a data project
- Subscriptions and resource groups: every Databricks workspace on Azure lives inside a subscription. Knowing this hierarchy helps separate development and production.
- Identity: access to storage and the workspace goes through Entra ID. Many pipeline permission errors start here.
- Storage: the Lakehouse sits on Azure storage accounts. Understanding redundancy and access tiers helps decide where each piece of data should live.
- Cost: a cluster left running costs money. The cost management part of the exam changes how you size compute.
How to study
The free Microsoft Learn material covers the whole exam. My advice is to study with the portal open: create a resource group, apply a policy, grant a permission with RBAC and then delete everything. Seeing the concept work sticks better than reading.
The next step
After AZ-900, the natural path for data engineering is DP-750 for Azure Databricks, which I took next. The cloud basics made the governance and deployment topics of that exam much easier.
See my other projects on the portfolio of Gabriel Henrique, Data Engineer. I'm Gabriel Henrique, and I write about Azure, Databricks and data engineering.