DeepGraph

DORA for GitHub projects

DORA is four numbers that show how often and how safely your team ships code. DeepGraph reads them from the pull requests, deployments, workflow runs, releases, and tags of your GitHub projects, so you see them next to the commits behind them.

Twelve weeks of a sample web shop: the four numbers, the chips of where it ships, and its deploys. Bars above the line worked, bars below it failed. Point at a name or a week to read more.

How it works

  • Connect GitHub. Sign in with GitHub and install the DeepGraph GitHub App on the projects you choose.
  • Approve access. The app asks to read pull requests, deployments, and workflow runs.
  • Open a report. At the top, switch from Code to DORA.
  • Check what counts. Chips beside the dates show where your projects ship. DeepGraph picks production for you, and you can change it.

What you can see

Deploys

How many times new code reached production in the dates you chose. This shows how often your team ships.

Lead time

How long a change takes from its first commit to production. Half of the changes got there faster. A shorter time means work reaches people sooner.

Failed deploys

The share of deploys that failed in production. A failure that is retried until it works counts once.

Time to restore

After a failed deploy, how long until the next good one. Half of the failures were fixed faster.

Hover over each name to read what it means. Under the numbers, a chart shows each day, week, or month: deploys that worked rise above a line, and failed ones hang below it. A dash means there is nothing to count yet, not zero.

For people and teams

DORA works for a folder, a project, a person, or a team. A person's numbers count the deploys that shipped their merged pull requests, and a team counts everyone in it. In the DORA view, the list of people shows each person's and team's deploys over time and their four numbers, and you can sort by any of them. Use this to see how work flows, not to rank people.

People of a sample web shop with their deploys over twelve weeks and their four numbers.

Choose what counts as production

Chips show deployments, workflows, releases, tags, and pull requests merged into a branch, each with its number of deploys. A filled chip counts as production.

Where a sample web shop ships. The filled chips count as production; staging does not.

DeepGraph's first guess skips names that read as staging, test, or preview. It prefers deployments and deploy workflows, then releases, then tags. With none of these, pull requests merged into the main branch count.

On your own project or folder, click a chip to count it or not, or open the gear to see every choice. A folder's choice applies to every project in it, and you can go back to DeepGraph's guess at any time. Anyone else can click a chip to look at one place only; nothing is saved.

What it doesn't do

  • It works with GitHub only. GitLab, Bitbucket, projects on your computer, and public projects added by their link have no DORA view.
  • Failed deploys and Time to restore need a place that records failures, such as GitHub deployments or workflow runs. Releases, tags, and merged pull requests only say that something shipped.
  • The first deploys can take a little while to show after you connect.

Questions

Why is my DORA tab empty?

Hover over the tab to see why. You may need to connect GitHub, or DeepGraph has not read any pull requests or deploys yet. See the FAQ.

Do I need to change how we deploy?

No. DeepGraph reads what GitHub already records. You only choose which places count as production.

What do the four numbers mean in general?

Read what DORA measures, in plain words.

Get started

Connect a GitHub project with the Getting started guide, or see plans and pricing. DORA sits next to project insights and teams. See all features.