Saving Package Configuration
Data Fusion supports publishing selected Data Integration (DI) configurations to the application package once pipeline construction is complete. Publishing seeds these configurations into the package as authoritative, application-level definitions.
During pipeline development, configurations such as source and target collections are created and refined directly on the Data Fusion canvas. Once these configurations are validated and intended for reuse, they can be published to the package to transition them from pipeline-local state to managed package metadata.
Published configurations are designed to support iterative pipeline development and production deployment. You can tune source, transform, and target configurations during development, then publish publish the finalized definitions so they can be reused consistently when the application is deployed to QA, UAT, or Production environments.
This separation allows interactive pipeline development to remain flexible, while ensuring that finalized configurations are treated as intentional, durable, and suitable for governed execution in production.
Publishing configurations is scoped to the application and the current environment. Publishing does not automatically make configurations portable across environments. Instead, it promotes finalized pipeline configurations into managed package metadata within the application.
Publishing to the package is the first step toward portability. Because packages are versioned and portable, published configurations can subsequently be promoted by publishing the package itself (for example, to a Git repository) and then deployed into downstream environments such as QA, UAT, or Production.
This model establishes a controlled workflow for evolving Data Fusion pipelines: configurations are iterated locally during development, published to the package once finalized, and then promoted through standard package deployment mechanisms.
Guidance
As a general rule:
- Keep configurations pipeline-local while iterating, validating assumptions, or experimenting.
- Publish configurations to the package once they represent intentional, reusable, and production-ready definitions within the application.
Publishing is optional and should be used deliberately to balance flexibility during development with stability and governance as pipelines mature.
Save Pipeline Configuration Values to the Package
After validating a pipeline configuration, you can sync selected configuration values to the application package so they can be reused, governed, and deployed across environments.
Use the Publish Pipeline Configurations feature to promote pipeline configuration values from the current environment to the package configuration. Publishing persists the selected configuration values in the package so they become part of the application’s versioned metadata.
Before you begin
- A Data Fusion pipeline is configured and valid.
- At least one Source or Target Collection contains configurable properties.
- You have permission to update the application package.
Steps
Open the Data Integration canvas.
In the upper-right corner of the canvas, click Save pipeline configurations.
The Save pipeline configurations dialog opens and displays configuration fields along with their current environment values and package values.
In the left panel, select the Source Collection Config or Target Collection Config you want to review.
Review the configuration fields listed in the table.
The table displays:
- Configuration Field – the pipeline configuration parameter.
- Environment Value – the value currently configured in the environment.
- Package Value – the value currently stored in the application package.
(Optional) Filter the configuration fields using the filter menu in the table header. Available options include:
- Show All – display all configuration fields.
- Show Differences – display fields where the environment value differs from the package value.
- Show Matches – display fields where the values match.
- Show only value in package – display fields currently stored in the package.
- Show only not in package – display fields that are not yet stored in the package.
Select the configuration fields you want to update using the checkboxes.
Click Copy Environment to Package to copy the selected environment values into the application package.
(Optional) Click Remove from Package to remove selected configuration values from the package.
Click Save to package to persist the changes, or click Save to package & close to save the changes and close the dialog.
After saving and publishing pipeline configuration values, the pipeline configuration is automatically recomputed from the updated package values.
Result
Selected configuration fields are marked for synchronization. Only selected fields are updated when you click Save to package. These values become part of the package configuration and can be deployed consistently across environments.
Saving pipeline configurations helps ensure that validated pipeline settings are preserved and included in the application package.
Publishing Pipeline Configuration (optional)
After saving pipeline configuration values to the package, open the GitHub menu and select Publish Changes to commit the updates to the connected repository. For more information about pushing local changes to a shared repository so collaborators can access and iterate on them, see the topic Version Control with GitHub listed under See Also.
Saving pipeline configurations operates at the pipeline level. The dialog aggregates configurable fields from nodes in the Data Fusion pipeline, such as Source Collection and Target Collection configurations. You can promote selected environment values for multiple nodes in a single operation so that the pipeline configuration can be versioned and reused across environments.
Verify the updated configuration
After syncing:
Package metadata is updated.
Pipeline values are recomputed from the package.
Synced fields are now governed and version-controlled.
Unselected fields remain pipeline-local and editable.
You have successfully promoted selected pipeline configuration values into managed package metadata, enabling consistent behavior across redeployments while preserving flexibility during pipeline development.