Data Connections & OneLake Catalog
Before you transform data, you need to find it. Learn how to discover data with OneLake catalog and Real-Time hub, and create connections to external sources.
How does data get into Fabric?
Think of data connections like plumbing in a building.
Before water (data) reaches your tap (lakehouse or warehouse), someone has to lay the pipes (connections) from the water main (source systems). Some pipes are permanent — they bring data from your company’s databases every night. Others are temporary — you hook up a garden hose to test something.
In Fabric, you discover what data exists using the OneLake catalog (like a building directory that shows where every pipe goes), and you connect to external sources using data connections (the plumbing itself).
OneLake catalog: Find data across your organisation
The OneLake catalog is your starting point for data discovery. Instead of asking colleagues “where is the sales data?”, you search the catalog.
What the catalog shows
| Information | Description |
|---|---|
| Item name and type | Lakehouse, warehouse, semantic model, dataflow, etc. |
| Workspace | Which workspace the item lives in |
| Endorsement | Whether the item is Promoted or Certified (trust signals) |
| Sensitivity labels | Classification level (Confidential, Internal, Public, etc.) |
| Description | Owner-provided context about the item |
| Lineage | What upstream and downstream items are connected |
Endorsement badges
Endorsement is how organisations signal data quality:
| Badge | Meaning | Who Can Apply |
|---|---|---|
| Promoted | Recommended by the workspace owner — ready for use | Workspace Members and above |
| Certified | Verified by a designated certifier — meets organisational standards | Only users granted certification permissions by the Fabric admin |
| None | No endorsement — use with caution | — |
Exam tip: Promoted vs Certified
The exam tests whether you know who can apply each endorsement. Promoted can be applied by any user with write permission on the item (Contributors, Members, Admins). Certified requires special permissions configured by the Fabric admin in the admin portal — only designated users or groups can certify items. A certified item carries more trust than a promoted one.
Real-Time hub: Discover streaming sources
The Real-Time hub complements the OneLake catalog for streaming data:
- Browse available streams — Event Hubs, Azure IoT Hub, Change Data Capture, custom streams
- Preview data — see sample events before connecting
- Connect to Eventhouse — create a direct link from a stream to an Eventhouse database
- Set up alerts — trigger actions when specific patterns appear in the stream
Scenario: Dr. Sarah discovers patient monitoring data
Dr. Sarah at Pacific Health Network opens the Real-Time hub and discovers a stream of patient vital signs from ICU monitors, published by the hospital’s IoT team. She previews the data — heart rate, blood pressure, oxygen saturation, timestamped every 5 seconds.
She connects the stream to an Eventhouse database with one click. Within minutes, she can query the live vitals data with KQL and build a real-time clinical dashboard.
Without the Real-Time hub, she would have needed to coordinate with the IoT team, exchange connection strings, and configure ingestion manually.
Creating data connections
Data connections in Fabric are centralised, reusable, and shareable.
Connection types
| Connection Type | Use Case | Authentication |
|---|---|---|
| Cloud connections | Azure SQL, Azure Data Lake, Dataverse, Snowflake, Google BigQuery | OAuth, service principal, connection string |
| On-premises data gateway | SQL Server, Oracle, file shares behind a firewall | Gateway cluster + credentials |
| OneLake shortcuts | Reference data in ADLS Gen2, S3, GCS without copying | Storage account keys, SAS tokens, service principals |
Creating a connection
- Go to Settings → Manage connections and gateways (or create inline within a pipeline/dataflow)
- Select the connection type (e.g., Azure SQL Database)
- Provide the server and database details
- Configure authentication (OAuth 2.0, SQL auth, service principal)
- Test the connection and save
- The connection is now reusable across pipelines, dataflows, and notebooks in the workspace
Scenario: James connects to 15 client databases
James at Summit Consulting sets up data connections for each client’s source database. He creates named connections like client-a-sql-prod and client-b-dataverse so his team can reference them by name in pipelines and dataflows.
Because connections are workspace-level resources, he can control who has access to each client’s credentials. A consultant working on Client A never sees Client B’s connection details.
James at Summit Consulting wants to mark a client's warehouse as trustworthy for the broader analytics team. He is a workspace Admin. Which endorsement level can he apply?
Raj at Atlas Capital needs to connect Fabric to an on-premises SQL Server database behind the company firewall. The database contains compliance data that cannot be moved to the cloud. Which connection method should he use?
Next up: Shortcuts & OneLake Integration — reference external data without copying it, and integrate Eventhouse data with the rest of Fabric.