New deployment method for Fabric Accelerator

Recently I tested the new deployment method for the Fabric Accelerator. Which is the metadata driven framework accelerator solution for Microsoft Fabric developed by Benny Austin. Because it is a popular solution I decided to share my results in this post.

Install with the new deployment method for Fabric accelerator

I looked to deploy fabric accelerator “as is” into a new workspace. In order to do this I first imported the fabric-accelerator and elt-framework GitHub repositories into my own account.

You need both GitHub repositories deploy the solution. First fabric-accelerator to deploy the service connection and Fabric items, then elt-framework to deploy all the items into the newly deployed SQL database in Fabric.

I then started going through the guide to deploy the Fabric Accelerator. Including configuring the secrets and changing the relevant environment variables. I then went to the Actions tab in my GitHub repository and ran the Deploy Fabric Accelerator workflow. Which completed as below.

Once completed, I went into Microsoft Fabric and checked that it had deployed all the items into my workspace.

I then looked to do the database deployment with the elt-framework repository. Which requires the details on the newly deployed SQL database in Fabric in the new workspace.

I had issues with the recommended connection string format during the deployment. However, the below format that I shared in a previous post about performing CI/CD for SQL database in Fabric works fine.

Server=tcp:{PROVIDED SERVER NAME IN FABRIC}.database.fabric.microsoft.com,1433;Initial Catalog={PROVIDED CATALOG/DATABASE NAME};MultipleActiveResultSets=False;Encrypt=True;TrustServerCertificate=False;Connection Timeout=30;Authentication=Active Directory Service Principal; User Id={SERVICE PRINCIPAL CLIENT ID}; Password={SERVICE PRINCIPAL SECRET}

Afterwards, the deployment was complete. From there I had various options. Including the ability to explore the solution or start working with it immediately.

Key points about new deployment method for Fabric Accelerator

Some key points I must highlight about the new deployment method after multiple installations:

• You can only deploy with GitHub only at this moment in time.
• Make sure you specify the correct Entra Object ID of Entra user or Security Group. Otherwise you will need Fabric admin rights to resolve some issues. Plus, the Fabric CLI to resolve connection issues.
• Your SQL database in Fabric quota applies to deleted workspaces as well. If testing with multiple installations, you need to permanently delete past workspaces as well.
• Adjusting the details about the data pipelines directly in the GitHub Actions workflow YAML file can cause issues. I recommend changing the details about them locally with Visual Studio Code. So that if encounter any issues can work with GitHub Copilot in Visual Studio Code to resolve issues.
• Be aware of the potential connection string format issue that I mentioned above.

Final words

I like the new deployment method for Fabric Accelerator. Because it makes the deployment experience a lot easier. Plus, it introduces new ways to customize the solution in a more modular way.

With this in mind, I recommend testing the new deployment method yourself.