Test Dashboard — Azure DevOps Extension

A testing dashboard for Azure DevOps focused on Test Plans, Test Cases, and execution results. It ships two contributions:

Test Dashboard hub — a full-page experience under the Test Plans area with four tabs:
- Dashboard — a sprint-scoped analytics view (see below).
- Test Plans — every plan in the project with state, area/iteration paths, and dates.
- Test Cases — browse cases by plan → suite, with state and automation status.
- Execution Results — overall pass/fail/blocked summary plus the most recent test runs.
Test Execution Summary widget — an at-a-glance dashboard tile showing the pass rate and outcome breakdown across recent runs.

Built with React + the official azure-devops-ui Formula design system and the azure-devops-extension-sdk / azure-devops-extension-api REST clients.

Sprint dashboard

The Dashboard tab has dependent Release → Sprint dropdowns that default to the current sprint (the one whose start/finish dates contain today). Six widgets re-query whenever the selection changes:

Test Execution Outcome — donut + stacked bar of passed/failed/blocked/N-A/not-run.
Test Execution Outcome by User Story — outcome per requirement-based suite's user story.
Open Bugs by Priority — donut grouped by priority.
Open Bugs by Assignee & State — stacked bars per assignee, segmented by state.
Bugs Root Cause Category — bars grouped by root-cause field.
Open Bugs — Blocker & High Priority — count + actionable list of P1/P2 / critical-high bugs.

Assumptions (adjust in code if your process differs)

Release/Sprint = the iteration tree. Top-level iteration nodes are treated as Releases and their children as Sprints (WorkItemDataService.getIterationTree).
"Open bug" = a Bug whose state is not Closed, Done, or Removed (CLOSED_STATES in WorkItemDataService).
Root Cause is read from whichever bug field ref-name contains RootCause / Root Cause (ROOT_CAUSE_HINTS); falls back to Unspecified.
Blocker & high = priority 1/2 or severity Critical/High (isBlockerOrHigh).
Test outcomes are scoped by plan iteration — plans whose iteration path is the selected sprint or a descendant (TestDataService.getPlansForIteration). "By user story" reads requirement-based suites and their linked work items.

Project layout

src/
  common/        Shared helpers (SDK init, async hook, status/empty views, outcome bar, table)
  services/      TestDataService — thin wrapper over the Test Plan + Test REST clients
  hub/           Full-page hub (Hub.tsx + components/)
  widget/        Dashboard widget (Widget.tsx)
vss-extension.json   Extension manifest (contributions, scopes, files)
webpack.config.js    One bundle per contribution iframe

Prerequisites

Node.js 18+
An Azure DevOps publisher — create one at https://marketplace.visualstudio.com/manage
The packaging CLI is installed as a dev dependency (tfx-cli)

Setup

npm install

Then set your publisher id in vss-extension.json (replace your-publisher-id).

Develop

npm run build:dev   # one-off development build into ./dist
npm run watch       # rebuild on change
npm run lint        # type-check only (tsc --noEmit)

Package & publish

npm run package     # builds production bundles and produces a .vsix in ./dist

Then either:

Upload the generated .vsix at https://marketplace.visualstudio.com/manage, or

Publish from the CLI:

npx tfx extension publish --vsix ./dist/*.vsix --token <PAT>

Finally, share the extension with your organization and install it from the Marketplace. After installing:

Open Test Plans → Test Dashboard for the hub.
On any team Dashboard, click Edit → Add widget → Test Execution Summary.

Required scopes

Scope	Why
`vso.test`	Read test plans, suites, points, runs, results
`vso.work`	Read work-item fields on test cases
`vso.project`	Resolve the current project context

Azure DevOps Server (on-prem) compatibility

The extension runs on Azure DevOps Services and Azure DevOps Server 2022.1 (on-prem). All REST/base URLs are resolved from the collection at runtime via the SDK location service (src/common/baseUrl.ts) instead of hard-coding dev.azure.com, so work-item links, AI comments, wiki publishing, and the current-iteration lookup all use your server's collection URL.

To install on Server: build the .vsix, then upload it via Collection Settings → Extensions → Manage extensions → Upload, and install it into the collection. Notes:

Requires Server 2022.1 (REST API 7.1). On older servers the wiki/comment calls (api-version=7.1 / 7.1-preview.3) may need lowering.
The AI review (Claude) calls the Anthropic API directly from the browser — the server must allow outbound HTTPS to api.anthropic.com, or leave the Claude key unset to skip it.

Notes

img/logo.png is a placeholder — replace it with your own 128×128 icon before publishing.
The widget has no configuration UI; it summarizes the last 50 test runs in the project. Add a config contribution if you want per-tile filtering.
Data volume: the Execution Results tab and the widget read recent runs (cheap, pre-aggregated). Per-plan point aggregation (getPlanOutcomeSummary) walks every suite and is intended for on-demand use.

Sprint Dashboard

T-Dashboard