Del via


Dashboard limits

This article describes the limits that apply to AI/BI dashboards, including pages, datasets, widgets, visualizations, and subscriptions.

Dashboard structure limits

The following limits apply to the overall structure and components of AI/BI dashboards:

Limit Maximum value Description
Pages per dashboard 15 pages A canvas can contain up to fifteen pages. New dashboards start with a single page.
Datasets per dashboard 100 datasets Define up to 100 datasets per dashboard. Datasets can be created from SQL queries, Unity Catalog tables or views, or uploaded files.
Widgets per dashboard 100 widgets per page Dashboards can hold up to 100 widgets per page. Widgets include visualizations, filters, text boxes, and images.

Filter limits

The following limits apply to dropdown and multiselect filters on AI/BI dashboards:

Limit Maximum value Description
Dropdown and multiselect filter values 100,000 values Dropdown and multiselect filters can render up to 100,000 distinct values. If the dataset contains more than 100,000 values, any values beyond that limit won’t appear in the list but users can still enter them manually. However, the search doesn't return matches for values that exceed the 100,000-value rendering limit.

Visualization rendering limits

The following limits apply to how much data can be rendered in dashboard visualizations:

Visualization type Maximum rows Notes
Most visualizations 10,000 rows To optimize performance, most charts can only render 10K rows on the canvas. Otherwise, visualizations can be truncated. This applies to area, bar, line, scatter, pie, heatmap, and other chart visualizations.
Table visualizations 100,000 rows Tables can display up to 100K rows before truncation.

Note

These row limits apply to the rendering of visualizations on the canvas. For information about dataset size thresholds and performance optimizations, see Dataset performance thresholds.

Dataset performance thresholds

While there's no hard limit on dataset size, Azure Databricks applies different performance optimizations based on dataset result size:

Dataset size Optimization behavior
Small datasets (≤ 100,000 rows or ≤ 100MB) The dataset result is pulled to the client, and visualization-specific filtering and aggregation are performed in the browser. Only the dataset query appears in the query history.
Large datasets (> 100,000 rows or > 100MB) The dataset query text is wrapped in a SQL WITH clause, and visualization-specific filtering and aggregation is performed in a query on the backend. The visualization query appears in the query history.

For optimal dashboard performance, keep your datasets small by using explicit column selection, WHERE clauses, and parameters to scope your SQL queries. See Dataset optimization and caching.

Dashboard subscription limits

Dashboard subscription emails enforce a 9MB combined attachment size limit that applies across all attached files. The following table describes individual limits.

File type Limit Behavior when limit exceeded
All email attachments combined (PDF, DesktopImage, CSV, TSV, and Excel) 9MB combined See below for specific behaviors based on which files exceed the limit.
Excel attachments 100,000 rows Rows beyond 100,000 are not included in the attachment.

Subscription size limit behavior

The following describes the expected behavior when the combined file size exceeds the 9MB limit:

  • If the PDF file is greater than 9MB: The subscription email doesn't include the PDF attachment or any images. It includes a note that says the dashboard has exceeded the size limit and shows the actual file size of the current dashboard.
  • If the combined file size is greater than 9MB: Only the PDF is attached to the email. The inline message includes a link to the dashboard but no image.
  • If tabular attachments are dropped due to the size limit: The email body includes the following message: "Attachments exceeded the 9 MB email limit and were excluded. Open the dashboard to download full results."

For more information about dashboard subscriptions, see Manage scheduled dashboard updates and subscriptions.