A set of status indicator lights, in the form of a stoplight, controlled by a Raspberry Pi, to monitor the current state of a continuous build system running Jenkins.
It could certainly be adapted to control other status indicators, and to monitor other build systems. This repository is for the software that monitors Jenkins and controlling the lights.
Once configured, the stoplight sits on a desk or prominent place and shows the real-time status of your most important (and separately your less important) continuous integration builds, 24/7 non-stop. No PC is required. The Raspberry Pi controller connects via wifi or wired ethernet connection to monitor the state of your build system.
See What Do All the Lights Mean? for an explanation of the status values.
To build your own extreme feedback device, you'll need a light and a Raspberry Pi to control it. My specific products ordered:
- A USB Traffic Light from cleware.de. This was the most expensive piece. I chose this model because (a) it's well built and had the look that I wanted, and (b) at the time I specifically wanted to learn to control devices via USB using Java. This could definitely build custom built for much cheaper, like using the Raspberry's GPIO to control individual LEDs or relays.
- A Raspberry Pi with Enclosure from amazon.com. These Vilros kits are the best value I've seen because they come with a complete kit, everything you need (including power supply) except an SD card.
- 8GB MicroSD Card, also from amazon.com. You can also get one with NOOBS already installed to save a step.
With this configuration, the stoplight runs on 5v USB power supplied by the Raspberry Pi, so no additional power supply is needed.
You also need a continuous integration server running Jenkins. The rest of this document will explain how to set up a Raspberry Pi and deploy the code from this repository to it, so monitor Jenkins and reflect its status.
The controller service built by this repository runs as a Java process, so the only hard requirement is that the Raspberry Pi have a Java VM. Most Raspberry Pis run some flavor of Linux, but it could technically be Windows 10 IoT or any other Raspberry OS.
The exact steps I used to set up the initial device are recorded in the Raspberry Pi Stoplight Installation guide, from installing the operating system to setting up the Linux service to auto-start when booted.
This project can be built and tested on any development machine (any platform supported by Gradle) and then deployed to a Raspberry Pi.
Compile and create a single distributable Jar that includes all dependencies using:
./gradlew fatJar
That will create build/libs/jenkins_xfd-all.jar, which can be deployed to a Raspberry Pi with USB light controls.
The complete Jar file can be simply copied to the Raspberry Pi (or test VM!) manually, usually from
build/libs/jenkins_xfd-all.jar to /usr/share/java/ on the target machine. However, for iterative
development it may be more convenient to configure Gradle to deploy directly to your target host.
TODO: We need to have some build properties that specify where the remote device is and how to connect and authenticate, but currently it's hardcoded in build.gradle.
First, set up passwordless access to the pi@stoplight machine (or whatever username@hostname it's called on your network), with something like cat ~/.ssh/id_rsa.pub | ssh pi@stoplight 'cat >> .ssh/authorized_keys'.
Then check the remotes section of build.gradle. It should look something like:
remotes {
stoplightServer {
host = 'stoplight'
user = 'pi'
identity = file('~/.ssh/id_rsa')
}
}
Specifying the remote user and host, as well as the local identity file.
Second, run the build to deploy to this target.
./gradlew deploy
Note: This build task is often hanging on the client side on Windows machines, even though it successfully deploys.
No special plugins are required in the Jenkins installation. The controller uses the RESTful endpoints natively exposed by Jenkins. However the How to Configure Jenkins page explains how to designate High and Medium priority builds.