Clean Rooms

The Clean Rooms page tracks activity for Databricks Clean Rooms — secure environments where multiple organizations collaborate on shared data without exposing their underlying datasets. Clean room notebook runs and collaboration events can be difficult to monitor, especially across collaborators. LakeSentry surfaces notebook run counts, durations, and collaboration events to help teams understand clean room usage.

Clean rooms overview

The top of the page shows headline metrics:

Metric	What it shows
Active Rooms	Number of clean rooms that have not been deleted
Total Runs	Total notebook executions across all clean rooms
Successful	Number of successful notebook runs
Collaborators	Total number of distinct collaborating organizations

Clean room list

The Overview tab shows all tracked clean rooms in a table, along with a bar chart of notebook runs by room (top 10):

Column	What it shows
Clean Room	Name and truncated central clean room ID
Status	Active or Deleted
Collaborators	Number of participating organizations
Notebook Runs	Total number of notebook executions, with a badge showing failed runs if any
Total Duration	Cumulative wall-clock time of all notebook runs
Last Activity	When the most recent event occurred, along with the event type

Sorting

The table defaults to sorting by Notebook Runs (descending). Sortable columns include:

• Clean Room — Alphabetical by name
• Collaborators — By collaborator count
• Notebook Runs — By total execution count

Daily activity

The Daily Activity tab shows per-day, per-room activity metrics for the selected time range. You can filter by a specific clean room using the room_id parameter.

Column	What it shows
Date	Day of recorded activity
Clean Room	Name of the clean room
Runs	Number of completed runs, with failed runs shown separately if any
Assets	Number of asset updates
Collaborators	Number of distinct collaborators active on that day
Avg Duration	Average notebook run duration for the day

Tracked event types

LakeSentry ingests the following event types from the system.access.clean_room_events system table:

Event type	What it records
CLEAN_ROOM_CREATED	When the clean room was created and by which collaborator
CLEAN_ROOM_DELETED	When the clean room was deleted
RUN_NOTEBOOK_STARTED	A notebook execution was initiated — includes which collaborator started it
RUN_NOTEBOOK_COMPLETED	A notebook execution finished — includes duration, status, and output schema
CLEAN_ROOM_ASSETS_UPDATED	Assets (tables, notebooks) were added, modified, or removed
ASSET_REVIEW_CREATED	An asset review was created — approvals, rejections, or auto-approvals
OUTPUT_SCHEMA_DELETED	An expired output schema was cleaned up (initiated by the system)

Each event records the initiating collaborator (identified by their alias — “creator”, “collaborator”, or a custom value) and the timestamp.

Note

Clean room events are tracked through the system.access.clean_room_events system table. Only one event record is emitted when collaborators share the same metastore. Cross-metastore collaborations generate separate records for each side.

Common workflows

Understanding clean room activity

1. Sort the clean room list by Notebook Runs (descending).
2. Review the bar chart to see which rooms have the most executions.
3. Check total duration to identify rooms with long-running notebooks.
4. Review failed run counts — failed runs still consume compute time.

Monitoring daily trends

1. Switch to the Daily Activity tab.
2. Review runs, asset updates, and collaborator counts per day.
3. Look at average run durations to spot performance changes over time.

Investigating failed runs

1. In the Overview tab, look for rooms with failed run badges.
2. Switch to Daily Activity to see when failures occurred.
3. Note the duration — long-running failures are expensive.
4. Compare successful vs. failed run counts to identify reliability issues.

Tip

If a clean room has a low notebook success rate, work with the collaborator who authored the notebook to investigate. Failed runs consume compute but produce no results — reducing failure rates directly reduces waste.

Next steps

• Cost Explorer — See clean room compute costs in the broader spending context
• Cost Attribution & Confidence Tiers — How costs are attributed to teams
• Audit Log — Full audit trail of clean room operations