Data Governance
Databricks Masks SQL Text by Default, With Privileged Readback
Databricks said on August 26 that query text is now masked by default in `system.query.history`, the Query History API, the List Queries API, and SQL-bearing audit parameters. Account admins and members of `databricks_pii_access` can read unmasked text, so the change narrows default exposure rather than deleting the underlying statements.
Citation-ready: Databricks now masks SQL query text by default in named system-table and API surfaces, while account admins and members of the `databricks_pii_access` group can read the unmasked statements.
What happened and why it matters
No. It changes what ordinary readers receive from named history and audit surfaces; privileged users can recover the text, and teams must still keep secrets and personal data out of queries and adjacent telemetry.
Official Databricks product release notes
Primary reference: Databricks August 2026 product release notes. Kaleido Field checked the event date and the article's attributed facts against this source.
| Source date | August 26, 2026 |
|---|---|
| Checked by Kaleido Field | August 31, 2026, 09:04 CST |
| Source function | current data-governance analysis separating default redaction, covered system tables and APIs, privileged access, query execution, retention, staged rollout, and broader sensitive telemetry |
Redaction changes the common read path
A user reading system query history or the named APIs now receives a redaction marker instead of the statement text. SQL definitions in specified audit request parameters are omitted rather than returned to the default reader.
That lowers casual exposure when statements contain customer names, filters, literals, or business logic. It does not make queries a safe place for secrets.
Privileged readback needs its own audit
Account admins can create the named group and add members who need unmasked access for debugging, incident response, or governance. Membership therefore becomes a sensitive entitlement.
Teams should record the approver, purpose, member, start and expiry, queries accessed, exports, alerts, and removal receipt, then review other logs, tags, notebook output, dashboards, and downstream observability separately.
Evidence boundary
Official product facts: affected fields, redaction marker, omitted audit parameters, privileged group, admin-grant path, and staged rollout warning. Not established: deletion of underlying statements, masking of every log or connector, retroactive coverage, protection from privileged misuse, retention changes, or compliance for a particular dataset.
FAQ
Which fields are masked?
The release note names `statement_text`, API `query_text`, and SQL-bearing keys in specified audit request parameters.
Who can see the original text?
Account admins and members of the `databricks_pii_access` account group.
Does rollout happen at once?
Databricks warns that staged releases can take a week or more to reach an account.