Azure DevOps

1. Overview

Azure DevOps (ADO) Boards is the system of record for work: what is wanted, what is wrong, who is doing it, and how far it has got. Code, pull requests and releases live on GitHub. The join between the two is deliberate and small: a work item id in a branch name and a pull request title, a pull request link on the work item, and a release name written back when the change ships.

This module describes how the project uses Boards. It is written for a developer or an AI session that has to create, move or query a work item and wants to do it the way everything else here is done. The rules are few and they are load-bearing: automation in the software factory reads states, fields and parent links and acts on them, so a work item that is shaped differently is one the automation cannot see.

Boards holds work. GitHub holds change. Do not duplicate one in the other: a pull request body does not restate the story, and a story does not carry a diff. The two vocabularies that govern the join are Work Item Tags and GitHub Labels; both are closed, and this module does not reopen them.

2. Pages

  • Work Item Process — the inherited Agile process, the five work item types and when each is used, the states and the lifecycle rule that moves an item from New to Closed as code merges and releases roll out, and the fields every item carries.

  • Additional Fields — the fields the project adds to the inherited process and how it uses the standard Activity field: Released In, Detected In and Target Environment, what writes each and what reads it.

  • Goals — the Goal work item: a dated outcome that cuts across streams, its fields, states and links, and how it relates to the goal tag and the Control Tower.

  • Usage Patterns — how to create a story, a bug, a task and a release task; acceptance criteria and the testing tasks that follow from them; linking pull requests; the standing queries; and what the factory writes on an item so you can read its trail.

  • Work Item Tags — the closed, faceted tag vocabulary and the test a candidate tag must pass.

  • GitHub Labels — the eight labels and why the set is small.

  • Development Workflow — branches, pull requests and releases on the GitHub side.

  • The Software Factory component on this site documents the automation that reads and writes work items.