dash-control 0.1.1

Databricks Control Center — live system-table dashboard for platform, governance, cost, users, jobs, and queries

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

Panel	What it shows
Health	DBU today, active users, failed jobs (last 24 h), total table count, 7-day DBU sparkline
Cost	Daily spend by SKU, top clusters, top jobs, top users by DBU, burn-rate projection
Users	Top users by query count & tables accessed, inactive users, permission changes
Catalog	Table inventory, tables by schema, stale tables, column count distribution, most-accessed tables
Jobs	Success rate, top failures, longest runs, daily run volume trend
Queries	Slowest queries, most expensive queries, top users, error summary
Governance	Tables without owners, PII columns, access anomalies, schema changes
Custom	Run any SQL against system tables; define custom panels in ControlCenterConfig

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

Package	Purpose
dash-dq	60+ data quality checks
dash-synthetic	Synthetic data generation
dash-observe	Freshness, volume & schema monitoring
dash-gov	Table/column lineage + role classification
dash-ontology	Auto-inferred business ontology from lineage
dash-control	Control Center — this package

License

Apache 2.0