Learning Objectives
By the end of this chapter, you will be able to:
- Configure Spark workspace settings, including starter pools, custom pools, runtimes, environments, and high-concurrency sessions
- Configure domain workspace settings, assign workspaces to domains and subdomains, and apply row-level, column-level, object-level, and folder/file-level access controls
- Configure OneLake workspace settings, including shortcut caching, default formats, and dynamic data masking
- Configure Dataflows Gen2 workspace settings, including workspace identity, staging, fast copy, and sensitivity labels
- Configure version control through Git integration and apply Promoted and Certified endorsements
- Implement SQL database projects and integrate Microsoft Fabric audit logs through Microsoft Purview
- Create and configure deployment pipelines with deployment rules and parameters, and configure OneLake security
- Implement workspace-level access controls and decide between a Dataflow Gen2, a pipeline, and a notebook for a given task
Executive Summary
- Microsoft Fabric organizes work into workspaces backed by an F SKU capacity. Workspace-level settings control Spark behavior, OneLake defaults, Dataflows Gen2 staging, domain membership, and Git wiring.
- Security in Fabric is layered. Workspace roles (Admin, Member, Contributor, Viewer) grant broad rights. Item permissions and OneLake data access roles refine access at the artifact and folder level. Row-level, column-level, object-level, and dynamic data masking rules apply at the storage engine.
- Lifecycle management uses Git integration for branching and Fabric deployment pipelines for stage promotion across development, test, and production workspaces. Deployment rules and parameters rebind connections, capacities, and configuration values between stages.
- Orchestration is a deliberate choice. Pipelines orchestrate, notebooks transform with code, Dataflows Gen2 transform with Power Query. Schedules and Activator-driven event triggers fire those orchestrations.
Assumptions
- The reader has working Microsoft Fabric tenant access with at least one F SKU capacity assigned.
- Microsoft Entra ID is used as the identity provider. References to AAD or Azure AD are not used; the correct name is Microsoft Entra ID.
- Code samples use fictional names like
fab-ws-finance-prod,lh_sales, andwh_financefor consistency. - Power BI items continue to use the term semantic model; the older term dataset is not used here.