Near real-time analytics on Azure Cosmos DB with Microsoft Fabric mirroring
summary
Mirroring copies Azure Cosmos DB (plus Azure SQL DB and Snowflake today) into Microsoft OneLake as open-source Delta format, creating an initial snapshot and then continuous incremental replication with no ETL and no impact to Cosmos DB request units. The system uses an internal CDC mechanism (not change feed or analytical store) and shows updates in fabric quickly — demo moves 7,500 → 8,000 rows with replication visible in <30 seconds. Replication compute isn’t charged and mirror storage is free up to allocated fabric capacity.
AI-written summary · about
chapters (7)
related talks
48mAccelerate insights with real-time Azure SQL data and Microsoft Fabric
25mAzure SQL MI & Fabric Fusion: A speed-session
51mA Deep Dive into Microsoft Fabric Data Warehouse
44mUnify your data with OneLake and Microsoft Fabric
17mAzure Cosmos DB in AI Workflows: From Knowledge Retrieval to Microsoft Fabric Integrations
47mData analytics for the era of AI
50mAzure SQL DB Data Portability : Mirroring, CDC, Export/Import and DataSync
20mThe What and the Why of Microsoft Fabric Real-time Analytics
