llms.txt

Dashboards

Assemble visualizations into interactive dashboards with shared filters and layouts.

What Are Dashboards?

Dashboards are collections of visualizations arranged in a flexible grid layout. They bring together multiple charts, tables, KPI cards, and interactive controls into a single view, giving you a comprehensive picture of your data at a glance.

A well-designed dashboard answers a specific set of questions. An operations dashboard might show error rates, active users, and page load times. A product dashboard might display feature adoption, user flows, and conversion metrics. An executive dashboard might present high-level KPIs with trend lines and geographic breakdowns.

Shared Parameters (Dashboard Filters)

One of the most powerful features of dashboards is shared parameters — filters that apply across all visualizations on the dashboard at once. Instead of setting a date range on each chart individually, you define a date range parameter at the dashboard level and all connected visualizations update together.

Dashboard parameters use the same type system as query parameters: text, numbers, dates, booleans, and arrays. Common shared parameters include:

  • Date range — Filter all visualizations to the same time period
  • Application — Show data for a specific app when you have multiple applications
  • Environment — Switch between production, staging, or development data
  • Region — Focus on a specific geographic area
Add a Date Picker visualization to the top of your dashboard to give users an interactive way to change the date range. The date picker feeds its selected value into the dashboard's shared parameters, so all charts update automatically when the user picks a new date.

Adding and Arranging Visualizations

Building a dashboard involves selecting visualizations and placing them in a grid layout. Each visualization you add to a dashboard becomes an instance — a reference to the original visualization with optional configuration overrides specific to that dashboard.

The grid layout lets you control:

  • Position — Where the visualization appears in the grid (row and column)
  • Size — How many grid cells the visualization spans (width and height)
  • Order — The visual stacking and reading order of visualizations

You can place the same visualization on multiple dashboards. Each instance is independent — changing the configuration on one dashboard does not affect how the visualization appears on other dashboards.

Per-Instance Configuration Overrides

When you add a visualization to a dashboard, you can override parts of its configuration without changing the original visualization definition. This is useful when you want the same chart to appear differently in different contexts.

Common overrides include:

  • Parameter bindings — Connect the visualization's query parameters to the dashboard's shared filters
  • Color schemes — Adapt chart colors to match a dashboard's visual theme
  • Axis labels and ranges — Customize presentation for a specific audience
  • Titles and descriptions — Provide context that is specific to this dashboard's purpose
Configuration overrides merge with the base visualization's settings. You only need to specify the properties you want to change — everything else is inherited from the original visualization.

Dashboard Display Components

Dashboards are rendered using a set of purpose-built display components:

ComponentPurpose
Dashboard ViewerThe top-level component that loads a dashboard and manages the shared parameter state. This is what you embed when you want to display a dashboard.
Dashboard DisplayRenders the dashboard layout with all of its visualization instances arranged according to their grid positions.
Dashboard GridThe grid layout engine that positions visualization cards based on their configured row, column, width, and height.

Typical Workflow: Building a Dashboard

  1. Plan your questions. Decide what questions this dashboard should answer. An operations dashboard might need to answer: "Are there errors right now?", "How many active users?", "Which pages are slowest?"
  2. Create or select queries. For each question, find an existing query or create a new one. Design queries with parameters for date ranges and filters.
  3. Create visualizations. For each query, create a visualization with the appropriate chart type. A trend question gets a line chart, a comparison gets a bar chart, a single metric gets a number card.
  4. Create the dashboard. Give it a name, key, and description. Define shared parameters like a date range that all visualizations will use.
  5. Add visualizations. Place each visualization on the dashboard, configure its grid position and size, and bind its query parameters to the dashboard's shared parameters.
  6. Add interactive controls. Place a Date Picker or Label visualization for user-facing filters and section headers.
  7. Set permissions. Share the dashboard with the appropriate teams by granting group access.

Sharing and Permissions

Dashboards have their own permissions, separate from the permissions on the visualizations and queries they contain. This gives you fine-grained control:

  • View — See the dashboard and interact with its filters
  • Edit — Add, remove, and rearrange visualizations; change dashboard settings
  • Manage — Control who else can access the dashboard
Dashboard permissions are layered. A user needs access to the dashboard itself, plus access to each underlying visualization and query, to see all the data. If a user lacks permission to a specific query, the visualization that depends on it will not show results — but the rest of the dashboard will still work normally.
Use the key field to give dashboards stable, memorable identifiers like ops-overview or product-metrics. Keys are easier to reference than numeric IDs and remain stable even if the dashboard is recreated.