Azure Data Factory is Microsoft's cloud service for building, scheduling and monitoring data pipelines that move and transform data between sources. It replaces manual, script-based integration work with a managed, visual pipeline you can automate, version and monitor from one place. Microsoft Fabric includes its own Data Factory, built on much of the same pipeline engine, for teams already working inside a Fabric workspace.

What Azure Data Factory does

A Data Factory pipeline is a sequence of activities. The most common is Copy Data, which reads from a source dataset and writes to a sink dataset – a dataset points at a specific source or destination, such as a container path or a database table. A pipeline can also run a stored procedure, call an Azure Function, trigger a Databricks notebook, or run a mapping data flow for row-level transformation logic.

Three pieces make a pipeline run on its own:

• Datasets point at a specific source or destination, such as a container path or a database table.

• Linked services hold the connection details a dataset needs, such as a server name and authentication method.

• Triggers decide when a pipeline runs: on a schedule, on a recurring window, or in response to an event such as a file landing in storage.

An integration runtime is the compute that actually moves the data. The Azure IR handles cloud-to-cloud movement. A self-hosted IR, installed on a machine inside your network, is what lets a cloud pipeline reach an on-premises SQL Server or file share.

A concrete pipeline example

A pipeline that copies a daily sales CSV file from Azure Blob Storage into an Azure SQL Database table looks like this, simplified from the JSON Data Factory generates:

{
  "name": "CopyDailySalesPipeline",
  "properties": {
    "activities": [
      {
        "name": "CopySalesCsvToSql",
        "type": "Copy",
        "inputs": [
          { "referenceName": "BlobSalesCsv", "type": "DatasetReference" }
        ],
        "outputs": [
          { "referenceName": "SqlSalesTable", "type": "DatasetReference" }
        ],
        "typeProperties": {
          "source": { "type": "DelimitedTextSource" },
          "sink": { "type": "AzureSqlSink" }
        }
      }
    ]
  }
}

A schedule trigger runs it every morning:

{
  "name": "DailyAt6amTrigger",
  "properties": {
    "type": "ScheduleTrigger",
    "typeProperties": {
      "recurrence": {
        "frequency": "Day",
        "interval": 1,
        "startTime": "2026-01-01T06:00:00Z",
        "timeZone": "UTC"
      }
    },
    "pipelines": [
      {
        "pipelineReference": {
          "referenceName": "CopyDailySalesPipeline",
          "type": "PipelineReference"
        }
      }
    ]
  }
}

In the Data Factory Studio, you build both pieces visually: drag a Copy activity onto the canvas, point its source and sink at the two datasets, then attach a schedule trigger under the pipeline's Add trigger menu. The JSON above is what that visual editor writes underneath. When the data needs reshaping on the way, not just copying, a pipeline can run a mapping data flow instead of a plain Copy activity. For lighter, Power Query-style transforms inside Power BI itself, see our guide to Power BI dataflows.

When to choose Data Factory in Fabric instead

Azure Data Factory is still a supported, actively developed service, and existing pipelines do not need to move anywhere. Choosing Data Factory in Fabric instead makes sense when:

• Your pipelines feed a Fabric Lakehouse or Warehouse, and you want them in the same workspace as those items instead of managed as a separate Azure resource.

• Your team already pays for Fabric capacity and wants pipeline and dataflow work billed from that capacity rather than as a separate Azure service.

• You want Dataflows Gen2, notebooks, pipelines and Power BI reports sitting in one workspace with shared permissions.

Azure Data Factory remains the better fit when your pipelines feed Azure services outside Fabric, or when existing CI/CD pipelines and infrastructure-as-code already target it and migrating has no clear benefit. On-premises sources are not a reason to stay: Azure Data Factory reaches them through a self-hosted integration runtime, and Fabric pipelines reach them through the on-premises data gateway. For the full feature-by-feature comparison, see our Data Factory showdown: Fabric vs. Azure and why choose Microsoft Fabric.

FAQs

Is Azure Data Factory still available now that Microsoft Fabric exists?

Yes. Azure Data Factory and Data Factory in Fabric are both current Microsoft products. Fabric did not replace or deprecate Azure Data Factory.

Do existing Azure Data Factory pipelines need to move to Fabric?

No. There is no forced migration. Moving a pipeline is a deliberate choice you make when the reasons above apply to your workspace.

Sources

• Introduction to Azure Data Factory - Microsoft Learn