Brigade Docs

Brigade: Event-driven scripting for Kubernetes.

Developer Guide

Developer Guide

This document explains how to get started developing Brigade.

Brigade is composed of numerous parts. The following represent the core components:

  • brigade-controller: The Kubernetes controller for delegating Brigade events
  • brigade-worker: The JavaScript runtime for executing brigade.js files. The controller spawns these, though you can run one directly as well.
  • brigade-api: The REST API server for user interfaces
  • brigade-project: The Helm chart for installing Brigade projects
  • brigade-vacuum: The stale build cleaner-upper (optional; enabled by default)
  • brig: The Brigade CLI
  • git-sidecar: The code that runs as a sidecar in cluster to fetch Git repositories (optional; enabled by default)

Additionally, there are several opt-in gateways that can be enabled via Helm chart values. These are:

  • Brigade GitHub App Gateway: The implementation of the GitHub App web hooks. It requires the controller.
  • Generic Gateway: A generic gateway offering flexibility to create Brigade events from webhooks originating from an arbitrary service/platform.
  • Container Registry Gateway: A gateway supporting container registry webhooks such as the ones emitted by DockerHub and ACR.

Read up on all the gateways above, as well as others, in the Gateways doc.

This document covers environment setup, how to run functional tests and development of brigade-controller and brigade-worker components.


  • Minikube
  • Docker
  • make

Clone the Repository In GOPATH

Building from source does not require code to be on your GOPATH since all builds are containerized by default, however, if you do have Go installed locally and wish (for instance) for your text editor or IDE’s Go support to work properly with this project, then follow these optional steps for cloning the Brigade repository into your GOPATH:

$ export GOPATH=$(go env GOPATH) # GOPATH is set to $HOME/go by default
$ export PATH=$GOPATH/bin:$PATH
$ mkdir -p $GOPATH/src/
$ git clone $GOPATH/src/
$ cd $GOPATH/src/

Note: this leaves you at the tip of master in the repository where active development is happening. You might prefer to checkout the most recent stable tag:

  • $ git checkout v1.0.0

After cloning the project locally, you should run this command to configure the remote:

$ git remote add fork<your GitHub username>/brigade

To push your changes to your fork, run:

$ git push --set-upstream fork <branch>

Working with Go Code

To run lint checks:

$ make lint

To format the Go files:

$ make format-go

To run the unit tests:

$ make test-unit

To re-run Go dependency resolution:

$ make dep

Working with JS Code (for the Brigade Worker)

To format the Javascript files:

$ make format-js

To run the tests:

$ make test-js

To re-run JS dependency resolution:

$ make yarn-install

(See Running the Brigade-Worker Locally below for live testing against a running instance.)

Building Source

To build all of the source, run this:

$ make build-all-images build-brig

To build just the Docker images, run:

$ make build-all-images

To build just the client binaries (for Mac, Linux, and Windows on amd64), run this:

$ make build-brig

Pushing Images

By default, built images are named using the following scheme: <component>:<version>. If you wish to push customized or experimental images you have built from source to a particular org on a particular Docker registry, this can be controlled with environment variables.

The following, for instance, will build images that can be pushed to the krancour org on Dockerhub (the registry that is implied when none is specified).

$ DOCKER_ORG=krancour make build-all-images

To build for the krancour org on a different registry, such as

$ DOCKER_ORG=krancour make build-all-images

Images built with names that specify registries and orgs for which you have write access can be pushed using make push-all-images. Note that the build-all-images target is a dependency for the push-all-images target, so the build and push processes can be accomplished together like so:

Note also that you must be logged into the registry in question before attempting this.

$ DOCKER_ORG=krancour make push-all-images

Minikube configuration

Start Minikube. Your addons should look like this:

$  minikube addons list
- addon-manager: enabled
- dashboard: disabled
- default-storageclass: enabled
- heapster: disabled
- ingress: enabled
- kube-dns: enabled
- registry: disabled
- registry-creds: disabled

Feel free to enable other addons, but the ones above are expected to be present for Brigade to operate.

For local development, you will want to point your Docker client to the Minikube Docker daemon:

$ eval $(minikube docker-env)

Running make build-all-images will build the Brigade images using the Minikube Docker daemon. The image tag will be derived from the git sha. You can verify this by running docker images and affirming these tagged images are listed.

Brigade charts are hosted in the separate brigadecore/charts repo, so we’ll need to add the corresponding Helm repo locally:

$ helm repo add brigade
"brigade" has been added to your repositories

If you just want to roll with the default chart values and let the Makefile set the image tags appropriately, simply run:

$ make helm-install

This will issue the appropriate command to create a new Brigade chart release on this cluster.

Note: helm init may be needed to get tiller up and running on the cluster, if not already started.

Note also: If you were specific about DOCKER_ORG and/or DOCKER_REGISTRY when building images, you should also be specific when running make helm-install. For instance:

$ DOCKER_ORG=krancour make build-all-images helm-install

During active development, the overall flow might then look like this:

$ # make code changes, commit
$ make build-all-images helm-upgrade
$ # (repeat)
$ # push to fork and create pull request

For finer-grained control over installation, do not use make helm-upgrade. Instead, you may opt to create a custom values.yaml file for the chart and set various values in addition to the latest image tags:

$ helm inspect values brigade/brigade > myvalues.yaml
$ open myvalues.yaml    # Change all `registry:` and `tag:` fields as appropriate

From here, you can install Brigade into Minikube using the Helm chart:

$ helm install -n brigade brigade/brigade -f myvalues.yaml

Don’t forget to also create a project. Check out projects to see how it’s done.

Running Brigade inside remote Kubernetes

Some developers use a remote Kubernetes instead of minikube.

To run a development version of Brigade inside of a remote Kubernetes, you will need to do two things:

  • Make sure you push your brigade docker images to a registry the cluster can access
  • Set the image when you do a helm install brigade/<chart> on the Brigade chart.

Running Brigade (brigade-controller) Locally (against Minikube)

Assuming you have Brigade installed (either on minikube or another cluster) and your $KUBECONFIG is pointing to that cluster, you can run brigade-controller locally.

$ ./bin/brigade-controller --kubeconfig $KUBECONFIG

(The default location for $KUBECONFIG on UNIX-like systems is $HOME/.kube.)

For the remainder of this document, we will assume that your local $KUBECONFIG is pointing to the correct cluster.

Running the Functional Tests

Once you have Brigade running in Minikube or a comparable alternative, you should be able to run the functional tests.

First, create a project that points to the brigadecore/empty-testbed project. The most flexible way of doing this is via the brig cli. Here we supply -x to forgo interactive prompts. All the defaults will therefore be set to the brigadecore/empty-testbed project.

 $ brig project create -x
Project ID: brigade-830c16d4aaf6f5490937ad719afd8490a5bcbef064d397411043ac

You can check this project configuration out via brig project get brigadecore/empty-testbed.

With this setup, you should be able to run make test-functional and see the tests run against your local Brigade binary.

Running the Brigade-Worker Locally

You can run the Brigade worker locally by cding into brigade-worker and running k brigade. Note that this will require you to set a number of environment variables. See brigade-worker/index.ts for the list of variables you will need to set.

Here is an example script for running a quick test against a locally running brigade worker.


export BRIGADE_EVENT_TYPE=quicktest
export BRIGADE_COMMIT_REF=master
export BRIGADE_PROJECT_ID=brigade-830c16d4aaf6f5490937ad719afd8490a5bcbef064d397411043ac
export BRIGADE_SCRIPT="$(pwd)/brigade.js"

cd ./brigade-worker
yarn start

You may change the variables above to point to the desired project.