Custom Reports: Configurable Widget Layout, Issue Dataset, and Date Trends
The Custom Reports builder gains a configurable dashboard layout (per-widget placement, size, and chart type), an Issue dataset for the inbuilt defect manager, automatic stacked bar/column charts when two dimensions are picked, and date-grouped trends like 'test cases executed per day'.
- Configurable widget layout on the project dashboard - Pinned report widgets are no longer fixed in a single grid slot. Each widget now has its own placement (row/column), size (width/height), and chart type, persisted via a new dashboard widget layout API. Customers can put 'Test Execution by Department' and 'Test Execution by Module' on separate full-width rows, resize a widget to span the dashboard, or switch its chart type without rebuilding the underlying report.
- Issue dataset - Inbuilt issues are now a fourth dataset alongside Test Cases, Test Plans, and Executions, gated to projects using the inbuilt defect manager. Pivot on standard fields (status, priority, created by, issue key, dates) or any defect custom field. Status and priority values come from the project's own
defect_field_options, so charts reflect each project's configured workflow.- Linked test case / suite / test plan dimensions - Within the Issue dataset, group by the test case, test suite, or test plan an issue is linked to — useful for tying defect counts back to coverage. Joins are added only when these dimensions are selected, so reports that don't need them avoid row fan-out.
- Stacked charts and per-widget chart type - Column and bar charts auto-stack when a report has two dimensions (e.g., issues per department × status). Each widget now has its own chart type selector — column, bar, line, or pie — and the choice is persisted in the report definition, so a single report can mix chart styles across widgets.
- Date dimensions and date trends - Date fields (
run_on on executions, created_at on every dataset) can now sit on the X-axis, grouped by day. Line charts are auto-selected when the primary dimension is a date, and date categories sort chronologically — so 'test cases executed per day' and similar trend reports work out of the box.

