Evergreen is a distribution of Jenkins we are working on that provides an easy to use and automatically upgrading experience. This year at the conference, there will be not just one, but two talks to present Evergreen to the Jenkins community:

Tyler will present the overall Jenkins Evergreen architecture, its inception and how this aims at making it much simpler for people to just use Jenkins to build their projects, without having to become Jenkins admins.

On the last conference day, during my own talk I will focus on the improved developer experience, and zoom into how we implemented some important features.

We will dig together into the Error Telemetry system put in place, allowing us to actually fix errors and warnings people see in production environments. How instances are automatically reporting errors to the Evergreen backend, and how we then centralize and analyze them using Sentry. We will explain how the Incrementals system allows developers a very short roundtrip, between a merged pull-request and a release we can push out to all instances. We will see concrete examples of issues we already fixed and released to Evergreen instances in just a few days after we opened an alpha version to the world.

I will show you how an instance starts up and gets upgraded by discussing with the backend it’s constantly connected to. How the backend knows what it should instruct an instance to download and install, or how we trigger an automated data snapshot.

You will obviously see a demo of all this showing in particular how Evergreen can already run on a Docker host, or on AWS (more environments to come!), using some of the so-called flavors for Jenkins Evergreen.