DataFlow and DataPipeline cross workspace

Hi DwijaKeerthi ,

This looks like a cross-workspace metadata resolution issue rather than a configuration mistake.

When a pipeline in Workspace2 references a Dataflow in Workspace1, Fabric allows you to select the artifact. However, parameter metadata does not always load across workspace boundaries. That is why you see:

“Failed to load parameters, please manually enter parameter details.”

 

Parameter auto-population depends on:

• Workspace-level permissions
• Capacity alignment
• Artifact type, Dataflow Gen2 vs classic
• Metadata context resolution across workspaces

Even if the dataflow is selectable, the parameter schema may not hydrate properly.

Things to verify

1. Permissions
   Ensure you have at least Contributor or Member access in Workspace1. Viewer is often not sufficient for metadata resolution.

2. Capacity
   Both workspaces should be on the same Fabric capacity. Cross-capacity references can block parameter introspection.

3. Dataflow type
   If this is Dataflow Gen2, cross-workspace parameter handling can be inconsistent depending on the build.

Workarounds

Option 1
Manually define the parameters in the pipeline and map them to the dataflow parameters. This works reliably even when auto-load fails.

Option 2
Place the pipeline in the same workspace as the dataflow. Parameter auto-binding typically works immediately in the same workspace context.

Option 3
Expose the output via Lakehouse or Warehouse and orchestrate at that layer instead of directly binding to the dataflow.

In short, this behavior is common in cross-workspace setups. Manual parameter entry is currently the most stable approach unless both artifacts live in the same workspace and capacity.

Thank you!
Proud to be a Super User!
📩 Need more help?
✔️ Don’t forget to Accept as Solution if this guidance worked for you.
💛 Your Like motivates me to keep helping