dash-control

Databricks Control Center — a live, 8-panel system-table dashboard for platform, governance, cost, and operations teams. One %pip install + one launch() call gives you an always-on view of your entire Databricks workspace.

%pip install dash-control
from dashcontrol import launch, ControlCenterConfig

config = ControlCenterConfig(
    date_range_days=30,
    catalogs=["main", "analytics"],
)
launch(config)

Panels

Configuration

from dashcontrol import ControlCenterConfig, CustomPanel

config = ControlCenterConfig(
    date_range_days=14,          # default look-back window (also adjustable in the UI)
    catalogs=["main"],           # limit catalog scope; empty = all catalogs
    panels=["health", "cost"],   # show only these panels
    row_limit=200,               # max rows per table result
    workspace_name="prod",       # displayed in the header
    custom_panels=[
        CustomPanel(
            name="My Query",
            sql="SELECT user_name, COUNT(*) AS cnt FROM system.access.audit GROUP BY 1",
            description="Custom audit rollup",
        )
    ],
)

Requirements

• Databricks Runtime 13.3 LTS or later (system tables must be enabled)
• Unity Catalog workspace
• ipywidgets (included as a dependency)

Architecture

Every panel is lazy-loaded — clicking Load runs the query; nothing executes on launch(). Queries target Databricks system tables (system.billing.*, system.access.audit, system.jobs.*, system.query.history, system.information_schema.*). All queries degrade gracefully: if a system table isn't available on your workspace tier, the panel shows an info banner instead of crashing.

Part of the DashLibs suite

License

Apache 2.0