Velo – DORA Metrics & AI Pipeline Intelligence for Azure DevOps

AI-Powered DORA Metrics, Pipeline Analytics, and Engineering Intelligence for Azure DevOps

Velo gives engineering managers and development teams instant visibility into how software is actually being delivered. It automatically tracks all five DORA metrics — including the 2026 Rework Rate — directly from your Azure DevOps pipelines and pull requests, with no manual tagging or custom scripts required. When you need to go deeper, the built-in Pipeline Intelligence Agent (powered by Microsoft Foundry + GPT-4o) answers questions about your delivery health in plain language, right inside Azure DevOps.

Built on Microsoft Foundry + Azure PaaS. Native Azure DevOps extension — one-click install, zero configuration.

Install from the Visual Studio Marketplace →

Screenshots

Features

DORA Metrics (all 5, including 2026 Rework Rate)

• Deployment Frequency — deployment-tagged pipeline runs ÷ 30-day window (falls back to all successful runs when no deployment-tagged pipelines are detected; dashboard shows "estimated" in that case)
• Lead Time for Changes — average time from PR merge to the first successful production deployment (in hours), based on ADO git.pullrequest.merged events linked to subsequent deployment-tagged runs. Falls back to the average successful build duration when fewer than 3 PR→deploy linkages are available in the window; in that case the dashboard shows an "approx." badge (IsLeadTimeApproximate=true).
• Change Failure Rate — failed deployment runs ÷ total deployment runs (falls back to all runs when no deployment-tagged pipelines exist; dashboard should show "estimated" in that case)
• Mean Time to Restore — average time from a failed deployment run's finish time to the next successful run's finish time, of the same pipeline (in hours) (falls back to all runs when no deployment-tagged pipelines exist; dashboard should show "estimated" in that case)
• Rework Rate (2026) — work-item state-transition churn: items that moved from a completed state back to an active state ÷ total completions, via ADO workitem.updated service hook events (shows "insufficient data" when no work-item events have been received)

Pipeline Intelligence Agent

• Conversational chat powered by Microsoft Foundry + GPT-4o
• Answers questions about pipeline bottlenecks, team health, and delivery trends
• Responses cached per pipeline fingerprint — no redundant AI calls
• Token-budgeted: 50,000 input tokens / org / day on the free tier

Azure DevOps Native

• Project Hub — full metrics dashboard + agent chat in the left nav
• Dashboard Widgets — compact DORA tiles for any ADO dashboard
• Velo@1 Pipeline Task — optional instrumentation step in YAML pipelines
• One-click install from the Visual Studio Marketplace; no infrastructure required for end users

Tech Stack

Architecture

VS Marketplace CDN          Azure PaaS (your subscription)
┌──────────────────┐        ┌─────────────────────────────┐
│  Angular 19      │◄──────►│  ASP.NET Core 9 API         │
│  Extension UI    │  Auth   │  Azure Web Apps       │
└──────────────────┘  B2C   │  WebhookController (ADO)    │
                            ├─────────────────────────────┤
Azure DevOps ──────────────►│  Azure SQL Serverless        │
 (service hooks)            │  Row-level security (RLS)    │
                            ├─────────────────────────────┤
                            │  Microsoft Foundry Agent     │
                            │  GPT-4o (cached by pipeline) │
                            └─────────────────────────────┘

Quick Start

Install from Marketplace

1. Go to your Azure DevOps organization → Organization Settings → Extensions
2. Search for Velo or install directly from the Visual Studio Marketplace
3. The Velo tab appears in every project's left navigation immediately

Deploy Your Own Backend (Enterprise / Self-Hosted)

See docs/extension/self-host-backend.md

Local Development

Prerequisites

Setup

# Clone
git clone https://github.com/divyeshg94/Ingenious.Velo.git
cd Ingenious.Velo

# Backend — restore, build, run
dotnet restore
dotnet build
cd src/Velo.Api && dotnet run

# Extension — separate terminal
cd src/Velo.Extension
npm install
npm start

Local Database

# Start SQL Server via Docker
docker run -e "ACCEPT_EULA=Y" -e "SA_PASSWORD=Velo@Dev123" \
  -p 1433:1433 --name velo-sql \
  -d mcr.microsoft.com/mssql/server:2022-latest

# Apply migrations (repeat for 002–015)
sqlcmd -S localhost -U sa -P Velo@Dev123 -d master \
  -i db/migrations/001_initial_schema.sql

Secrets

Use dotnet user-secrets — never commit connection strings:

cd src/Velo.Api
dotnet user-secrets set "ConnectionStrings:VeloDb" "Server=localhost;Database=VeloDev;..."

For the full dev environment walkthrough see docs/contributing/local-setup.md.

Repository Structure

Ingenious.Velo/
├── src/
│   ├── Velo.Extension/     # Angular 19 ADO Extension
│   ├── Velo.Api/           # ASP.NET Core 9 REST API
│   ├── Velo.Agent/         # Microsoft Foundry AI Agent
│   ├── Velo.SQL/           # EF Core DbContext + migrations
│   └── Velo.Shared/        # Shared models and contracts
├── infra/                  # Bicep IaC (dev/staging/prod)
├── db/migrations/          # SQL migrations (001–015)
└── docs/                   # Architecture, API, and dev docs

Roadmap

Implemented

Up Next

Future

Contributing

Contributions are welcome. Please read CONTRIBUTING.md first, then:

1. Fork the repo and create a branch: git checkout -b feature/your-feature
2. Follow the coding standards and testing guide
3. Ensure CI passes: dotnet test and npm test
4. Open a PR against master with a clear description of what and why

For larger changes, open an issue first to align on the approach.

Acknowledgements

• Microsoft Foundry / Azure AI Agents — agent orchestration
• Azure DevOps Extension SDK — extension host integration
• Angular, ASP.NET Core, EF Core — application frameworks

License

MIT — see LICENSE