Create a GitHub connection

Important

This feature is in Beta.

Create a GitHub connection in Catalog Explorer to store authentication credentials for Lakeflow Connect ingestion. Any user with the USE CONNECTION privilege on the connection can then create GitHub ingestion pipelines without needing direct access to the credentials.

Alternatively, admin users can create a connection and a pipeline at the same time in the data ingestion UI. See Create an ingestion pipeline.

For privilege requirements, see Connect to managed ingestion sources.

Prerequisites

Complete the source setup. Use the authentication details you obtain to create the connection.

Create a connection

  1. In the Azure Databricks workspace, click Data icon. Catalog > Create > Create a connection.
  2. On the Connection basics page of the Set up connection wizard, enter a unique Connection name.
  3. In the Connection type drop-down menu, select GitHub.
  4. (Optional) Add a comment.
  5. Click Next.
  6. On the Authentication page, enter the following credentials:
    • Client ID: The client ID from the source setup.
    • Client secret: The client secret from the source setup.
    • (Optional) Enterprise domain: The hostname of your GitHub Enterprise instance. Leave this field empty for GitHub.com. See Enterprise hosts.
  7. Click Authorize. You're redirected to the GitHub authorization page.
  8. Review and authorize all the requested scopes. You're redirected to the Azure Databricks workspace.
  9. Click Create connection.

Enterprise hosts

Use the Enterprise domain field to select the GitHub host for your connection:

GitHub host Enterprise domain REST API base URL
GitHub.com Leave empty. https://api.github.com
GitHub Enterprise Cloud on GHE.com <tenant>.ghe.com https://api.<tenant>.ghe.com
GitHub Enterprise Server (GHES) Your server hostname, such as github.example.com. https://github.example.com/api/v3

Create the OAuth app on the same GitHub instance that hosts your organization. Your workspace must have network access to the instance and its API host.

Next steps

Create an ingestion pipeline.