Problem

You are trying to use databricks.sdk.AccountClient in a notebook, with default authentication in your workspace, when you get a NotFound: Endpoint not found error message.

Example

Using the Databricks SDK account client to list all of your account workspaces generates the following error:

NotFound: Endpoint not found for /2.0/accounts/<your-account-id>/workspaces

Cause

Notebook native authentication is chosen by default when using the SDK inside a Databricks notebook, unless you have specified a different way to authenticate in the account client.

Notebook native authentication is not supported at the account level. It only authenticates at the workspace level. 

Solution

You must provide a valid way to authenticate at the account level before you can use databricks.sdk.AccountClient in a notebook.

You can use OAuth with a service principal as detailed in the Authorize unattended access to Databricks resources with a service principal using OAuth (AWS | Azure | GCP) documentation.

Once you have the service principal client id and the secret, you can authenticate the SDK at the account level by calling AccountClient in a notebook.

You need to replace the following values:

• <databricks-account-url> -  Your account URL depends on your cloud. 
  • AWS: https://accounts.cloud.databricks.com
  • Azure: https://accounts.azuredatabricks.net
  • GCP: https://accounts.gcp.databricks.com
• <your-account-id> - You can find your account ID (AWS | Azure | GCP) in the account console.
• <client-id>  and <sp-secret> - your service principal ID and secret from when you created the service principal.

a = AccountClient(
host="<databricks-account-url>",
account_id='<your-account-id>,
client_id='<client-id>',
client_secret='<sp-secret>'
)

For more information, review the Databricks SDK for Python authentication documentation.