Microsoft Purview Now Scans Fabric Lakehouse Sub-Items for Enhanced Data Governance

Microsoft Purview has achieved General Availability for scanning Microsoft Fabric Lakehouse sub-item metadata, marking a significant advancement in automated data governance capabilities. This enhancement allows organizations to automatically discover and catalog detailed metadata from within Fabric Lakehouse environments, addressing a critical pain point for data teams managing complex, distributed data estates.

The Governance Challenge This Solves

Data governance teams have long struggled with fragmented visibility across growing data landscapes. As organizations deploy multiple Lakehouses, warehouses, and data pipelines, maintaining an accurate inventory of what data exists, where it resides, and its structure becomes increasingly difficult. Traditional approaches often rely on manual documentation, which quickly becomes outdated and incomplete.

Microsoft Purview directly tackles this challenge by automatically scanning Fabric tenants and extracting comprehensive metadata into the Unified Catalog. The key innovation lies in distinguishing between two levels of metadata extraction:

Item-level metadata covers top-level workspace artifacts including Lakehouses, warehouses, notebooks, and pipelines. These are treated as single entities in Purview and inventoried automatically after scan completion.

Sub-item metadata represents the breakthrough capability. Purview can now scan tables (specifically Delta format) and files within Lakehouses, surfacing column-level details, data types, and structural information directly in the Unified Catalog. This transforms governance from knowing "a Lakehouse called Sales Gold exists" to understanding "that Lakehouse contains a Delta table called fact_orders with 14 columns including order_date (date) and revenue (decimal)."

Technical Implementation

The integration requires several configuration steps, though Microsoft has streamlined the process significantly:

1. Register Fabric tenant as a data source in the Purview Data Map
2. Create security group in Microsoft Entra ID and add your Purview Managed Identity (MSI) or service principal
3. Grant read-only Admin API access to the security group in the Fabric tenant admin portal
4. Enable detailed metadata responses in the Fabric Admin portal - this critical setting ensures sub-item scanning functions correctly
5. Configure and schedule scans scoped to all workspaces or targeted subsets

For organizations already using Azure's identity infrastructure, Managed Identity authentication support simplifies credential management. However, teams running multiple Fabric or Power BI scans simultaneously should be aware of potential rate limiting. Microsoft recommends staggering scans across different time windows rather than running them in parallel.

Practical Impact for Data Teams

Once scanned, metadata surfaces in Purview's Unified Catalog where teams can browse by source type, workspace, or Fabric experience, and search for specific assets by name, description, or other attributes. This transforms data discovery from a tribal knowledge exercise into a searchable, governed process.

For data stewards and consumers, the benefits are substantial:

• Data discoverability: Analysts and data scientists can find Lakehouse tables without chasing down the engineer who built the pipeline months ago
• Data contracts: Column-level metadata enables more precise data contracts between producers and consumers
• Onboarding: New team members can evaluate data assets before requesting access
• Trust building: Automated metadata extraction reduces reliance on outdated documentation

Getting Started

Organizations ready to implement this capability can begin through the Microsoft Purview Portal by navigating to Data Map. The Register Microsoft Fabric in Microsoft Purview documentation provides detailed setup instructions.

The General Availability announcement, published March 3, 2025, represents Microsoft's commitment to deepening Fabric-Purview integration. As data estates continue growing in complexity, automated metadata extraction at the sub-item level becomes increasingly critical for maintaining governance, compliance, and operational efficiency.

For organizations already invested in the Microsoft data ecosystem, this capability represents a significant step toward unified data governance without the overhead of manual documentation or multiple disconnected tools.