Event‑Based Triggers for Fabric Pipelines (e.g., SharePoint / OneDrive File Change Events)

[trishailendra]

Today, Microsoft Fabric Data Pipelines primarily support time‑based scheduling (hourly, daily, etc.) and manual triggers. However, many real‑world data integration scenarios are event‑driven, especially when source data resides in SharePoint Online, OneDrive, or Microsoft 365 storage.

A common enterprise scenario is:

• An Excel file is uploaded or modified in a SharePoint document library
• Downstream data pipelines should immediately process the updated data

Currently, achieving this requires:

• External orchestration tools (Power Automate, Logic Apps)
• REST API calls with Entra ID admin consent
• Custom governance approvals and additional operational complexity

This introduces friction, especially for organizations that want to keep all orchestration natively inside Microsoft Fabric.

Proposed Feature / Enhancement

Introduce native event‑based triggers for Microsoft Fabric Data Pipelines, similar to Azure Data Factory and other event‑driven platforms.

Example Trigger Types

• File Created
• File Modified
• File Deleted

Example Event Sources

• SharePoint Online document libraries
• OneDrive for Business
• Fabric OneLake (Files / Lakehouse folders)
• Future extensibility: Azure Blob Storage, ADLS Gen2
  

  Example Use Case

  When an Excel file is modified or replaced in a SharePoint folder, automatically trigger a Fabric pipeline to ingest, validate, and transform the data.

  Flow

  1. User updates Excel file in SharePoint
  2. Fabric detects file modification event
  3. Pipeline is triggered automatically
  4. Data is ingested into Lakehouse / Warehouse
  5. Downstream notebooks or semantic model refreshes run

  No external automation, REST calls, or Power Automate flows required.

  ---

  Suggested UX / Configuration

  • Add “Event Trigger” as a pipeline trigger type (alongside Schedule & Manual)
  • Allow users to:
    • Select source (SharePoint, OneLake, etc.)
    • Provide folder path and file pattern (e.g., *.xlsx)
    • Choose event type (Created / Modified)
  • Support basic debounce logic (e.g., ignore multiple changes within X minutes)
  • Optionally expose file metadata as pipeline parameters (file name, modified time, user)