Skip to main content

Governance Studio

Governance Studio is the DFS Pro workspace for repeatable data processing pipelines. Use it when a workflow needs governed input datasets, reusable methods, run history, and node-level execution evidence.

Typical uses:

  • prepare reviewed data before a BI report;
  • apply a reusable method to a validated dataset;
  • produce an output dataset for fusion or AI Agent workflows;
  • inspect run evidence after a processing failure;
  • standardize a data preparation workflow across sites.

Before you start​

Confirm:

  • dfs:read and dfs:write access;
  • DFS Pro package access;
  • at least one prepared dataset;
  • at least one published or draft method suitable for the workflow;
  • a named owner for the resulting output.

Open Governance Studio​

Go to:

Data Integration > Governance Studio

The list view shows pipelines, status, node count, update time, and delete action.

Create a pipeline​

  1. Select New Pipeline.
  2. Enter a pipeline name.
  3. Add a description.
  4. Save.
  5. Open the canvas.

Use a name that states the operational outcome.

Examples:

  • Facility sensor quality preparation
  • Work order and asset feature pipeline
  • Inspection evidence normalization
  • Energy meter reporting preparation

Add dataset nodes​

In the canvas:

  1. Select the Dataset Nodes palette.
  2. Search for the dataset.
  3. Drag the dataset to the canvas.
  4. Confirm the label in the node inspector.

Use validated datasets where possible. Draft datasets can be useful during early design, but production pipelines should have a stewarded input set.

Add method nodes​

  1. Select the Processing Nodes palette.
  2. Search for a method.
  3. Drag the method to the canvas.
  4. Connect dataset nodes to method nodes.
  5. Select the method node to review label and method details.

Use methods from the Method Library when the same data processing logic should be reused or versioned.

Save the pipeline​

Select Save after arranging nodes and edges.

Before saving for shared use, confirm:

  • each node label is understandable;
  • the flow direction is clear;
  • methods match the input dataset shape;
  • the owner can explain each processing step.

Validate the pipeline​

Select Validate before running.

Review:

  • validation status;
  • node count;
  • edge count;
  • method count;
  • sink count;
  • errors;
  • warnings.

Fix validation errors before running the pipeline. Treat warnings as review items for the pipeline owner.

Run the pipeline​

Select Run Pipeline when validation has passed or the owner accepts the remaining warnings.

During a run, the pipeline can show running, completed, or failed status. Use the run drawer to inspect execution history.

Inspect run history​

Open Runs to review previous pipeline runs.

Check:

  • run ID;
  • status;
  • start time;
  • duration;
  • triggering user when visible;
  • registered output datasets;
  • error message when present.

Use run history during troubleshooting and for evidence review.

Inspect node execution logs​

After a run, select a node and open its execution log.

Review:

  • node ID;
  • node label;
  • node status;
  • method key when present;
  • duration;
  • rows in;
  • rows out;
  • error message;
  • preview rows when available.

Node logs are useful when the full pipeline fails but only one dataset or method caused the issue.

Troubleshooting​

SymptomCheck
Dataset is missing from paletteConfirm dataset access, status, and tenant scope.
Method is missing from paletteConfirm method category, status, and permissions.
Validation failsReview missing edges, missing methods, invalid dataset references, and warnings.
Run failsOpen run history, inspect node logs, and fix the first failed node.
Output dataset is unavailableCheck registered outputs in run history and downstream dataset permissions.

Next step​

Use Audit and Metrics to review operational health after pipelines are running.