Databricks permissions management at scale

[lukaszmaron]

Working with permissions in Databricks at scale gets tricky pretty fast.

A few patterns that keep coming up:

• permissions spread across workspaces, catalogs, groups
• hard to answer simple questions like “who actually has access to X?”
• lots of manual querying / scripting to reconstruct the full picture

Even when audit logs are enabled, they don’t fully solve:

• current state vs historical access
• effective permissions (inherited via groups)
• cross-workspace visibility

Feels like there’s still a gap between how permissions are modeled and how easily they can be analyzed or audited in practice.

Checkout the article and the tool I've been working on 🙂

Full article: Medium 

GitHub repo: Repo

[KrisJohannesen]

Cool solution!

A lot of the auditing of what you have built will actually be available as part of the Governance Hub. It is built on the system tables
- It also contains a lot of other cool features that are more in the areas of Data Quality, Classification etc.

It is currently in Private Preview, so you need to contact your Databricks Account Team to enroll - but sounds like you might benefit from seeing it - and could have some cool ideas for the team. Highly recommend it!