Understanding the Databricks Cost Dashboard

Once data starts flowing from your connectors, the LakeSentry dashboard gives you a complete picture of your Databricks spend. This page walks through the main sections and how to navigate them.

The Overview page

The Overview page is your starting point. It shows a high-level summary of key metrics, cost trends, and top spenders across your entire Databricks environment.

Stat cards

At the top of the page, stat cards summarize key metrics at a glance:

• Total Cost — Current period spend with percent change from the previous period.
• Cost of Failure — Spend attributed to failed job runs, linking to waste-related insights.
• 30-Day Forecast — Projected spend range with trend direction and confidence level.
• Active Jobs — Count of work units with activity in the selected time range.
• Insights — Count of active insights (with critical count highlighted), linking to the full Insights queue.
• Attribution Quality — Percentage of spend that has been attributed to teams or owners.

Cost trends

Below the stat cards, the cost trend chart shows your total Databricks spend over time, with day, week, and month granularity options. In Breakdown cost mode, the chart splits into Databricks and Cloud cost layers. The chart also automatically detects and highlights anomalies — unusual spikes or drops appear as visual markers directly on the chart. Use this to spot trends — is spend growing? Are there periodic spikes? Did something change recently?

Top spenders

Below the trend chart, you’ll see the top cost drivers — the jobs, pipelines, SQL warehouses, clusters, and serving endpoints consuming the most budget. A cost category breakdown (jobs & pipelines, SQL warehouse, interactive, serving, overhead) lets you filter by compute type. Each entry shows cost, type, and activity metrics.

Click any item to expand its details or drill into its detail view.

Global controls

These controls appear in the header across all pages:

Time range selector

Controls the analysis window for all charts and tables on the page. Options include:

• Preset ranges: 24 hours, 7 days, 30 days, 90 days, 1 year, month to date, year to date, and all time
• Custom range: Pick any start and end date

The time range is always visible in the header so you know exactly what period you’re looking at.

Workspace filter

If you have multiple workspaces, use this multi-select filter to focus on specific ones. By default, all workspaces are included.

Organization filter

Filter by your organizational hierarchy (org unit, department, team). This is useful once you’ve set up attribution rules and org hierarchy mappings.

Tag filter

Filter by Databricks resource tags. Select specific tag keys and values to narrow down cost data to tagged resources.

Cost mode toggle

Switch between different cost views:

Mode	Shows
Total	Databricks + Cloud combined
Databricks Only	Databricks costs
Cloud Only	Cloud infrastructure costs
Breakdown	Show Databricks and Cloud separately

Main navigation sections

LakeSentry organizes features into main sidebar groups:

Investigate

• Overview — The dashboard described above: trends, top spenders, anomalies.
• Cost Explorer — Deep multi-dimensional cost analysis with 8 tabs: Breakdown, Compute Types, Principals, Tables, Comparison, Matrix, Trends, and Attribution. See Cost Explorer for details.
• Cost Changes — Period-over-period changes and new or discontinued cost drivers.

Optimize

• Insights — The queue of actionable findings: anomalies, waste detection, and optimization suggestions. Each insight includes evidence, estimated savings, and a path to action. See Insights & Actions.
• AI Agent — Preview cost investigation assistant.

Infrastructure

Detailed views for each type of Databricks resource:

• Jobs & Pipelines — Jobs and pipelines with per-run cost history. See Work Units.
• Clusters — Interactive and job cluster utilization metrics. See Compute.
• SQL — Query-level cost analysis for warehouses and serverless. See SQL Analysis.
• Serving — Model serving endpoint costs and traffic patterns. See Model Serving.
• Storage — Storage cost breakdown and trends. See Storage.

Manage

• Budgets — Track spend against budgets at the team, department, or workspace level. See Budgets.
• Commitments — Track and manage Databricks commit contracts.
• Mappings — Manage organizational hierarchy mappings.
• Cost Allocation — Configure allocation rules that map costs to teams and owners. See Cost Allocation.
• Tag Governance — Manage tag policies and compliance. See Tag Governance. This item is hidden for scoped non-admin users.

System

• Settings — Organization preferences, display options, and configuration.
• Audit Log — Record of all actions, approvals, and changes. See Audit Log.

Quick navigation

LakeSentry includes a command palette (press ⌘K or Ctrl+K) for quick navigation and search. Type the name of any page, resource, or entity to jump directly to it.

Tip

Use the command palette (⌘K / Ctrl+K) to quickly jump to any resource, page, or entity by name. This is the fastest way to navigate when you know what you’re looking for.

Typical investigation flow

Here’s how a typical cost investigation works using the dashboard:

1. Start on Overview — Notice a cost spike in the trend chart or an anomaly marker.
2. Identify the driver — Check top spenders to see which workspace, job, or user is responsible.
3. Drill into details — Click through to the specific resource (e.g., a job’s detail page) to see per-run cost history.
4. Check insights — Visit the Insights page to see if LakeSentry has already flagged the issue and suggested an action.
5. Take action — Approve the suggested optimization, or use the information to make changes in your Databricks environment directly.

Note

The investigation flow above follows the most common path, but you can start from any page. For example, if you’re focused on a specific team’s costs, start from Cost Explorer filtered by team rather than the Overview.

Data freshness

The Overview page shows data as fresh as the latest successful connector sync and Databricks system-table latency allow. If data seems stale, check connector health and Data Sync on the Connector tab in Settings. For details, see Data Freshness & Pipeline Status.

Next steps

• Cost Explorer — Deep-dive cost analysis with multi-dimensional breakdowns
• Insights & Actions — Review and act on optimization findings
• How LakeSentry Works — Understand the data pipeline behind the dashboard