Azure Databricks Notebook Sharing, Notebook Exporting and Notebook Clipboard copy download

[VJ3]

Hello,

I would like to know in which scenario Azure Databricks User would be able to download Notebook Command Output if Notebook Result Download is disabled. Do we know if Privilege user would be able to share sensitive information with non-privilege user from his/her workspace?

How can we ensure that User only should be able code and no notebook command output or notebook results with another user in same workspace or different workspace?

Thank you

Vijay

[Kaniz · Accepted solution]

Hi @VJ3, Certainly! Let’s delve into the details regarding Azure Databricks and how to manage notebook outputs and results.

Notebook Outputs and Results:

• When a cell is run in a notebook, it produces outputs such as tables, plots, or text results.
• By default, downloading results is enabled. Users can download cell results containing tabular output to their local machines.
• To toggle this setting, workspace administrators can manage the ability to download results from not....
• If you want to clear the notebook state and outputs, you can choose from the following options in the Run menu:
  • Clear all cell outputs: Clears the cell outputs. Useful for sharing a notebook without including any results.
  • Clear state: Clears the notebook state, including function and variable definitions, data, and imported libraries.
  • Clear state and outputs: Clears both cell outputs and the notebook state.
  • Clear state and run all: Clears the notebook state and starts a new run.

Sharing Sensitive Information:

• Privileged users should exercise caution when sharing notebooks with non-privileged users.
• If a privileged user runs a cell that contains sensitive information (e.g., credentials, proprietary data), the output of that cell will be visible to anyone with access to the notebook.
• To prevent accidental sharing of sensitive information, consider the following measures:
  • Review Outputs: Privileged users should carefully review the outputs before sharing notebooks. Clear any sensitive results.
  • Workspace Permissions: Ensure that non-privileged users have restricted access to sensitive notebooks.
  • Collaboration Policies: Establish guidelines within your organization to prevent inadvertent sharing of sensitive data.

Code-Only Sharing:

• To ensure that users can only share code without notebook outputs or results:
  • Educate users about the importance of reviewing and clearing outputs.
  • Set up workspace permissions to restrict access to sensitive notebooks.
  • Encourage users to follow best practices when sharing notebooks.

Remember that while Azure Databricks provides powerful collaboration features, responsible usage and security practices are essential to protect sensitive information.

[VJ3]

Thank you Kaniz. Can we disable the exporting of notebook except Source File? If yes, then how do we achieve is?

Also, we do not want to share the notebook which has any kind of notebook results, can we use spark.databricks.query.displayMaxRows and set it to zero so that zero results will be shared with another user?