Skip to main content

Dashboards

Dashboards provide customisable views of test data using a grid of configurable widget types. Create personal or shared dashboards for single-project monitoring or cross-project comparison.


Creating a Dashboard​

  1. Navigate to Dashboards and click Create Dashboard
  2. Enter a name and optional description
  3. Choose visibility: Public (visible to all users) or Private (only you)
  4. Choose a Product Lane Strategy:
    • Single Product Focus — all widgets share one product lane configured at the dashboard level. Changes to the global filter apply across all widgets simultaneously.
    • Multi-Product Flexibility — each widget independently chooses its own product lane. Use this for cross-product or cross-team dashboards.
  5. Click Add Widget to start building — choose a widget type, configure it, and add it to the grid

Dashboard layouts auto-save as you make changes. Rename a dashboard by clicking its title inline.


Dashboard Filters​

In a Single Product Focus dashboard, a global filter panel at the top applies across all widgets simultaneously. The filter state is preserved in the URL — bookmark or share specific views.

Core Filters​

Product & Type (required), Team, Browser, Device, Platform, Stage, Version

Advanced Filters​

Date Range, Build Range, Status Filters, Relation Filters

Each widget can also have local filter overrides. Click the widget's gear icon to configure per-widget settings.


Product Lane Configuration​

Understanding how product lanes are configured — and what happens when fields are left empty — is essential for getting accurate data in each widget.

How Product Lane Resolution Works​

Each widget resolves its data source in this priority order:

  1. Widget-level product lane — if the widget has its own product lane configured in the editor, that takes precedence over everything else. Only available in Multi-Product Flexibility dashboards — the Product Lane panel is hidden in Single Product Focus dashboards because all widgets inherit the dashboard-level filter.
  2. Dashboard-level query — if no widget-level override exists and the dashboard has a product lane selected (Single Product Focus), the widget inherits it.
  3. Additional lanes — a list of extra product lanes for side-by-side comparison. Can coexist with the primary lane above. Each entry fetches data independently.

If none of the above is configured, the widget renders empty — it has no data source to query.

Required vs Optional Fields​

FieldRequiredBehaviour when empty
Product✅ YesWidget will not fetch data
Type✅ YesWidget will not fetch data
TeamNoWildcard — matches builds from all teams
VersionNoWildcard — matches all versions
BrowserNoWildcard — matches all browsers
DeviceNoWildcard — matches all devices
PlatformNoWildcard — matches all platforms
StageNoWildcard — matches all stages
ExtrasNoWildcard — no extra-field filtering
info

Leaving an optional field empty is not the same as filtering for null — it means the field is excluded from the query entirely, so all values (including null, blank, and any non-null value) are included.

Single Lane​

The most common configuration. Specify a product + type combination, optionally narrowed by team, stage, version, etc. The widget fetches the N most recent builds matching that filter exactly.

Example: Product = Reporting, Type = API, Stage = Staging, Version = v1.2 → returns the last N builds for that exact combination.

When the filter matches exactly one distinct product lane combination, the widget operates in single lane mode.

:::caution Subset trap Leaving an optional field empty does not select a "parent" lane — it is a wildcard that matches all values. If you configure Product=UReport, Type=UI, Stage=Staging with Version left empty, and your data contains builds for v1.1, v1.2, and v1.3, the filter matches three distinct combinations and aggregation mode is triggered automatically. There is no way to target "all versions as a single combined lane" in single-lane mode once multiple versions exist in the data. Your options are:

  • Specify an explicit version to stay in single-lane mode
  • Accept aggregation mode (data pooled as Σ 0, Σ 1, … indices)
  • Use Break aggregation (Build History Bar only) to show each version as its own independent series
  • Use Single series (Build History Bar only) to show all matched builds as one flat chronological series :::

Aggregated Lanes (auto-detected)​

When you configure a product lane filter but leave one or more optional fields empty, the filter may match multiple distinct lane combinations in your data — for example, leaving Browser empty when you have Chrome, Firefox, and Safari builds all under the same product/type.

UReport detects this automatically when you apply a filter. If exactly one lane combination matches, the widget shows a normal single-lane view. If more than one matches, it switches to aggregated lanes mode — all matching sub-lanes are pooled together.

In aggregation mode:

  • The chart uses Σ 0, Σ 1, … labels on the X-axis, where each index represents the Nth most recent run across all matched sub-lanes.
  • Clicking a bar or data point opens an overlay listing each individual sub-lane build that contributed to that data point.
  • Some widget types are hidden — Summary Card widgets and Test Analysis Table do not support aggregation and are removed from the widget type dropdown. You must configure a single-lane filter to use those widget types.

Example: Product = Reporting, Type = API, Browser = (empty) — if Chrome, Firefox, and Safari builds all exist for this product/type, the widget automatically aggregates them.

Impact per widget:

  • Build History Bar — each bar represents combined counts across all matched sub-lanes for that run index
  • Status Pie / Triaged Pie — combined test counts from the latest run across all matched sub-lanes
  • Heatmap — each row pools tests from all matched sub-lanes for that run index
  • Treemap / Relations Group By — tests from the latest run across all matched sub-lanes, pooled before grouping
  • Trend Insights — trend computed from the aggregated pass rate across all matched sub-lanes
  • Top Tests Table — unaffected; queries by product/type regardless of sub-lane count

Multi-Lane​

Supported by widgets that show side-by-side comparisons (Build History Bar and Relations Group By). Add extra product lanes in the widget editor under Multiple Product Lanes. Each lane is fetched independently and rendered as a separate series.

Example: Lane 1 = Chrome, Lane 2 = Firefox, Lane 3 = Safari → three independent series on the same chart.

When multiple lanes are configured:

  • A Grid View toggle appears in the widget header — shows all lanes as small charts side-by-side.
  • A single view shows one lane at a time; use the grid view to see all simultaneously.

Aggregation Mode Options (Build History Bar only)​

When aggregation mode is detected (filter matches more than one lane), the Build History Bar widget editor shows a three-option selector:

OptionBehaviour
Aggregated (default)Builds are pooled into Σ 0, Σ 1, … run indices. Clicking a bar shows each sub-lane's contribution.
Break aggregationAuto-discovers each sub-lane on every load and renders them as independent series — a special form of Multi-Lane that never hardcodes the lane list. New versions or parameter values appear automatically without re-saving.
Single seriesAll matched builds across all sub-lanes are flattened into one chronological series, one bar per actual build, sorted by start date. X-axis switches to Start Date automatically. Use this when you want a single continuous history view regardless of which sub-lane each build came from.

The selector only appears when your filter matches more than one lane. For single-lane filters the widget always operates in single-lane mode and the selector is hidden.

:::tip When to use Single series Choose Single series when you want to see "all staging builds, regardless of version, as one flat timeline" — for example, to spot overall quality trends without caring which version each build belongs to. Note that build numbers may collide across sub-lanes, so the X-axis defaults to Start Date. :::


Widget Types​

Build History Bar​

Bar or line chart showing pass/fail/skip/triage counts across multiple builds over time.

What it reads: Up to N builds (configured via range) matching the product lane, plus all tests for each build. Falls back to build-level status counts if individual test data is unavailable.

X-axis options:

  • Build Number — sequential build number

  • Start Date — actual execution timestamp. Format is chosen automatically based on both the total number of builds and the time span they cover, so labels stay readable regardless of how dense the chart is:

    BuildsSpanFormatExample
    > 20anyMMM ddJun 15
    > 10 or span > 10 days—MMM dd, yyyyJun 15, 2026
    ≤ 10, span > 1 day—MMM dd HH:mmJun 15 14:30
    ≤ 10, same day—MMM dd HH:mm:ssJun 15 14:30:45

    The hover tooltip always shows the full date with year and seconds regardless of the axis label format.

Chart type toggle: Switch between Bar and Line at runtime. Line mode plots pass rate (0–100%) per build. Bar mode stacks pass/fail/skip/triage/outage counts. This is the same widget — the toggle is available at runtime in the widget header.

Product lane scenarios:

ScenarioBehaviour
Single laneOne series of bars/points, one per build.
Aggregated lanesEach bar represents combined counts across all matched sub-lanes for that run index (Σ 0, Σ 1, …). Clicking a bar opens an overlay showing each sub-lane's contribution. Use the aggregation mode selector to switch to Break aggregation or Single series.
Break aggregationLanes auto-discovered on every load, rendered as independent series (special form of multi-lane). Lane list updates automatically as new versions appear.
Single seriesAll matched builds flattened into one chronological series sorted by start date, one bar per actual build. X-axis switches to Start Date automatically.
Multi-laneMultiple independent series — one per manually configured lane. Grid view shows them side-by-side; single view shows one lane at a time.
No product laneWidget shows empty — no builds to display.

Analytics: When data is loaded, an Analytics overlay is available (chart icon in header). Shows pass-rate trend, build health distribution, and stability score.


Status Pie Chart​

Pie chart of test result distribution for the latest build only.

What it reads: Tests from the single most recent build matching the product lane filter.

Two modes (toggle in header):

  • Status view — slices for Pass / Fail / Skip / Triaged (KI) / Outage
  • Triage view — slices showing the breakdown by investigation cause (Defect, Flakiness, Environment Issue, etc.)

Product lane scenarios:

ScenarioBehaviour
Single laneShows status distribution of the latest build in that lane
Aggregated lanesCombines tests from the latest run of each matched sub-lane. Pie reflects the totals across all groups.
Multi-laneNot applicable — Status Pie does not support multi-lane. Configure separate widgets per lane.
No product laneWidget shows empty.

Triage view note: Issue type colours come from your Settings → Issue Types configuration. Up to 5 types are shown as individual slices; remaining types are grouped into "others".


Triaged Pie Chart​

Identical to Status Pie Chart but opens on the Triage view by default — slices show the breakdown by investigation cause (Defect, Flakiness, Environment Issue, etc.) rather than run status. All product lane scenarios and behaviour are the same as Status Pie Chart.


Summary Cards​

A compact card grid showing key metrics. A single widget can display three different card sets — switch between them at runtime using the selector at the top:

Build Summary​

Counts and buckets all builds in the configured range by pass rate:

BucketRange
🏆 Perfect100%
✅ Good90–99%
✅ Acceptable80–89%
⚠️ Warning60–79%
❌ Poor40–59%
🚫 Criticalbelow 40%

Also shows average pass rate with a trend indicator (↑ / ↓) comparing the first half to the second half of the build range.

Status Summary​

Aggregates test counts across all builds in the range (not just the latest). Shows executed total, pass, fail, skip, triaged, and outage. Also shows an "active" count (executed minus skipped/outaged) when relevant.

Triaged Summary​

Shows triage coverage across all builds: total triaged tests, new triage (first analysed in the current build) vs carried over (auto-matched from a previous triage). Breakdown by issue type using your configured issue type colours.

Product lane scenarios (applies to all three card types):

ScenarioBehaviour
Single laneMetrics computed from builds in that lane over the configured range
Aggregated lanesNot supported — Summary Card widgets are hidden from the widget type dropdown when your filter matches multiple lane combinations. Configure a single-lane product filter to use these widgets.
No product laneWidget shows empty.

Heatmap​

Colour-coded matrix: rows represent builds (newest at top), columns represent values of a relation dimension. Cell colour encodes pass rate — green (high) to red (low).

What it reads: N builds and all tests per build, then groups tests by the selected relation dimension (component, team, tag, or custom field).

Configuration:

  • Group By — select which relation dimension to use as columns (configured in widget editor; can be changed at runtime via the selector in the widget header)
  • Build range — how many builds (rows) to show

Product lane scenarios:

ScenarioBehaviour
Single laneOne row per build, one column per relation value
Aggregated lanesEach Σ index becomes a row; tests from all matched sub-lanes are pooled before grouping
No product laneWidget shows empty.

Clicking a cell opens a detail drawer listing the individual builds that contributed to that cell, with links to launch the test view for that specific build.


Treemap​

Hierarchical visualisation of test distribution across a relation dimension. Block size is proportional to test count; colour encodes pass rate.

What it reads: Tests from the latest build, grouped by the selected relation dimension.

Product lane scenarios:

ScenarioBehaviour
Single laneTests from the latest build in that lane
Aggregated lanesTests from the latest run of each matched sub-lane, pooled together
No product laneWidget shows empty.

Clicking a block navigates to the Launches view filtered to that relation value.


Relations Group By​

Stacked bar chart grouping tests from the latest build by a relation dimension. Each bar represents one relation value; segments show pass/fail/skip/triage breakdown.

Modes:

  • Horizontal bars
  • Vertical bars
  • Table — tabular view with pass rate column

Multi-lane support: Relations Group By supports Multiple Product Lanes (same as Build History Bar). Add extra product lanes in the widget editor to show side-by-side comparisons across lanes.

Product lane scenarios:

ScenarioBehaviour
Single laneGroups tests from the latest build
Aggregated lanesGroups tests from the latest run across all matched sub-lanes
Multi-laneMultiple independent series — one per configured lane
No product laneWidget shows empty.

Test Analysis Table​

Tabular analysis of test behaviour patterns across all builds in the configured range. Classifies each test as:

  • Stable — consistently passes across builds
  • Unstable — passes in some builds, fails in others (flaky)
  • Consistently failing — fails in most or all builds

Shows flakiness percentage, total runs, and fail count per test.

Product lane scenarios:

ScenarioBehaviour
Single laneAnalyses tests across the configured build range for that lane
Aggregated lanesNot supported — Test Analysis Table is hidden from the widget type dropdown when your filter matches multiple lane combinations. Configure a single-lane product filter to use this widget.
No product laneWidget shows empty.

Supports CSV export of the full table.


Top Tests Table​

Ranked table surfacing the most impactful tests. Unlike most widgets, this widget queries an analytics API over a time window rather than a build range.

Two modes (toggle at runtime):

ModeShows
Top FailuresTests that failed most frequently; columns: fail count, last failed date
Slowest TestsTests with the longest average duration; columns: avg duration, max duration, run count

Time window options: Last 7 days / 30 days / 90 days (configurable in the widget header at runtime).

Product lane scenarios:

ScenarioBehaviour
Single laneTop tests within that product/type (and any optional filters) over the time window
Aggregated lanesThe widget remains available and queries the analytics API using the configured product/type filter. Because the analytics API is time-window based (not build-pipeline based), results reflect all data matching the product/type regardless of multi-lane configuration.
No product laneWidget shows empty.

Trend Insights​

Detects anomalous pass-rate drops across recent builds using Z-score analysis.

What it reads: Pass rates across the last N builds (default 20). Computes a statistical baseline and flags builds where the pass rate deviated significantly.

Overlays:

  • Pass rate timeline with anomaly markers
  • Build duration trend — reveals if the suite is getting slower over time

Product lane scenarios:

ScenarioBehaviour
Single laneTrend analysis for that lane
Aggregated lanesTrend computed from the aggregated pass rate across all matched sub-lanes
No product laneWidget shows empty.

Section Divider​

Visual separator that organises dashboard content into logical sections.

  • Customisable title, colour, font size, and line style
  • Enables jump-to-section navigation when multiple dividers are present
  • Spans full dashboard width
  • Does not fetch any data — no product lane configuration needed

Widget Configuration​

Each widget has specific configuration options accessible via the gear icon:

SettingDescription
Widget TitleCustom label for the widget header
Product Lane SelectionWhich product lane(s) the widget queries. Only shown in Multi-Product Flexibility dashboards — in Single Product Focus dashboards, widgets inherit the dashboard-level filter and this setting is hidden. See Product Lane Configuration above.
Date RangeOverrides dashboard-level date range for this widget
Build RangeNumber of recent builds to fetch (default: 20)
Status FiltersWhich statuses to include in calculations
Relation FiltersPre-filter tests by relation dimension values before the widget processes them
Chart Type / Visual optionsWidget-specific display settings (axis label, direction, colours, layout)

Multi-Lane Comparison​

Widgets that support multi-lane (Build History Bar and Relations Group By) can compare data from multiple product lanes side-by-side:

  1. Open the widget editor
  2. Expand Multiple Product Lanes
  3. Click Add Product Lane to add additional data sources — configure each with a different filter combination
  4. Use arrow buttons to reorder lanes; edit or remove as needed

Use cases:

  • Cross-Browser — Chrome, Firefox, Safari lanes side-by-side
  • Team Performance — Frontend, Backend, QA team lanes
  • Environment — Dev, Staging, Production comparison

The Multiple Product Lanes section only appears for widgets that support it.


Sharing and Collaboration​

FeatureDescription
Public dashboardsVisible to all logged-in users
Private dashboardsOnly accessible to the creator
Clone dashboardsCreate a personal copy of any dashboard
URL sharingDirect links with embedded filter state — shareable and bookmarkable
PDF ExportFull-page PDF snapshot of the current dashboard view
CSV ExportRaw data export from the Test Analysis Table widget

Layout Management​

FeatureDescription
Drag-and-dropMove widgets by dragging their header
ResizeDrag the bottom-right corner
Grid snappingAutomatic alignment to grid
Auto-saveLayout changes save immediately
Section DividersAdd Section Divider widgets to group related widgets and enable jump-to-section navigation
Widget lockWidgets auto-lock (drag/resize disabled) while dashboard filters are active — clear filters to re-enable layout editing

Next: Notifications →