Environment Git integration and deployment pipeline

This article describes how to use Git integration and deployment pipelines for environment in Microsoft Fabric.

Git integration for Microsoft Fabric environment

Microsoft Fabric supports Git integration, allowing developers to backup, control versions, revert to previous stages, and collaborate on their work using Git branches.

Important

This feature is in preview.

Note

• Currently, Git supports only Libraries and Spark compute, including Spark runtime.
• The Git integration manages the staging state of the environment, ensuring that both published and saved configurations are synced with Git.
• To apply changes made in Git to the environment, they must be published. It's recommended to publish after updating the environment from Git to ensure the effectiveness of the configuration.
• For a code-first experience, you can use the Publish API of environment to publish changes from Git without using the UI.
• Each commit has as upper limit of 150 MB. Custom libraries larger than 150 MB aren't currently supported through Git.

Connect the Fabric workspace to an Azure DevOps repository

If you are the admin of a workspace, navigate to the workspace settings, and set up the connection in Source control section. To learn more, see Manage a workspace with Git.

After connecting, you can find items including the environments syncing with the repo.

Local representation of an environment in Git

In the item root folder, environments are organized with a Libraries folder that contains PublicLibraries and CustomLibraries sub-folders, along with the Setting folder.

Libraries

When you commit an environment to Git, the public library section is transformed into its YAML representation. Additionally, the custom library is committed along with its source file.

You can update the public library by editing the YAML representation. Just like the portal experience, you can specify a library from PyPI and conda. You can specify the library with expected version, a version range or without version. The system can help you determine a version that is compatible with other dependencies in your environment. To clear all the existing public libraries, delete the YAML file.

You can update the custom library by adding new files or deleting existing files directly.

Note

You can bring your own YAML file to manage the public library. The file name needs to be environment.yml to be recognized correctly by the system.

Spark compute

The Spark compute section is transformed into the YAML representation as well. Within this YAML file, you can switch the attached pool, fine-tune compute configurations, manage Spark properties, and select the desired Spark runtime.

Deployment pipeline for environment

Important

This feature is in preview.

Fabric’s deployment pipelines simplify the process of delivering modified content across different phases, such as moving from development to test. The automatic pipeline can include the environment items to stream the re-creation process.

You can set up a deployment pipeline by assigning the workspaces with different phases. Learn more on Get started with deployment pipelines.

You can find the deploying status after setting up the pipeline successfully. After clicking the Deploy button with environment selected, all contents of the environment are deployed to the destination workspaces. The status of the original environment is preserved in this process, meaning the published configurations stay in published state and require no extra publishing.

Important

• The custom pool is currently not supported in deployment pipeline. If the environment selects the custom pool, the configurations of Compute section in the destination environment are set with default values. In this case, the environments keep showing different versions even the deployment is done successfully.
• Using deployment rules to specify different pool in new workspace will be included in upcoming release.

Related content

• Create, configure, and use an environment in Microsoft Fabric.
• Manage a workspace with Git.
• Get started with deployment pipelines.