Microsoft Fabric Blog provides a technical overview of Mirroring for SQL Server in Microsoft Fabric, describing how near real-time data ingestion enables seamless analytics and business intelligence for enterprise workloads.

Mirroring for SQL Server in Microsoft Fabric (Generally Available)

In today’s AI-driven world, analytics platforms depend heavily on the quality and timeliness of data. Enterprises collect huge volumes of data across applications, databases, and data warehouses, but ingesting this data centrally for analytics and AI is complex and costly. Traditional databases use proprietary formats, making shortcuts impossible and requiring extract, transform, and load (ETL) operations that rarely yield real-time results.

Introducing Mirroring in Microsoft Fabric

Mirroring now offers a modern method for continuous, near-real-time data access and ingestion from any database or data warehouse into OneLake within Microsoft Fabric. This ensures immediate access to data changes, eliminating stale insights and improving analytics responsiveness.

Mirroring is now Generally Available for all in-market SQL Server releases from 2016 to 2022, with support for SQL Server 2025. The feature simplifies the process: after configuration in the Fabric Portal (connecting your SQL Server databases and choosing what to mirror), an initial snapshot is created in Fabric’s OneLake and kept up to date in near-real-time, reflecting transactions, table changes, and data updates.

Technical Capabilities

Real-Time Analytics Foundation
- Immediate synchronization of source SQL Server data into OneLake
- No complex setup or ETL transformations required
- User selects eligible tables or mirrors entire databases
Version Support and Architecture Differences
- SQL Server 2016-2022: Relies on Change Data Capture (CDC) to snapshot and replicate changes. Requires on-premises data gateway (OPDG) or VNet data gateway for connectivity.
- SQL Server 2025: Uses ‘change feed’ technology (same as Azure SQL mirroring). Arc Agent is required for outbound authentication.
Integration with Fabric Analytics and BI Tools
- Mirrored data in delta format is instantly available for:
  - Power BI (Direct Lake mode)
  - Data Warehouse
  - Data Engineering workflows
  - Lakehouse
  - KQL Database
  - Notebooks
  - Copilots

Advantages

Simplifies data movement without complex or costly ETL operations
Enables near real-time analytics
Facilitates seamless integration for business intelligence and data science
Works across hybrid environments (on-premises SQL Server and Azure SQL)

Setup & Documentation

Configure mirroring via the Fabric Portal, specifying database connection and table selection
Detailed step-by-step instructions available in Mirrored SQL Server documentation

Table: Key Differences Between SQL Server Sources And Mirroring Methods

Source	Change Capture	Arc Agent	Data Gateway
SQL Server 2016-2022	CDC	Not required	Required for writing
SQL Server 2025	Change Feed	Required	Used for control plane
Azure SQL	Change Feed	System identity	Required in private nw

Once data is mirrored, it is immediately consumable across Fabric experiences for advanced analytics, reporting, and machine learning workloads.

Resources

This post appeared first on “Microsoft Fabric Blog”. Read the entire article here