Developer Guide
This document explains how to get started developing Brigade.
Brigade is composed of numerous parts. The following represent the core components:
- apiserver: Brigade’s API server, where the majority of core Brigade logic lives
- scheduler: Schedules event workers and jobs on the underlying substrate
- observer: Observes (and records) state changes in workers and jobs and ultimately cleans them up from the underlying substrate
- worker: The default runtime for executing
brigade.js/.ts
files. - logger: Event log aggregator, with Linux and Windows variants
- brig: The Brigade CLI
- git-initializer: The code that runs as a sidecar to fetch Git repositories for vcs-enabled projects
This document covers environment setup, how to run tests and development of core components.
Prerequisites
- A local Kubernetes cluster, 1.16.0+. We recommend kind or minikube.
- Docker or Go 1.16+ and JS dependencies
- make
Clone the Repository
The first step to begin developing on Brigade is to clone the repo locally, if you haven’t already done so. First navigate to your preferred work directory and then issue the clone command:
# Clone via ssh:
$ git clone git@github.com:brigadecore/brigade.git
# Or, via https:
$ git clone https://github.com/brigadecore/brigade.git
After cloning the project locally, you should run this command to configure the remote:
$ git remote add fork https://github.com/<your GitHub username>/brigade
To push your changes to your fork, run:
$ git push --set-upstream fork <branch>
Containerized Development Environment
To ensure a consistent development environment for all contributors, Brigade relies heavily on Docker containers as sandboxes for all development activities including dependency resolution, executing tests, or running a development server.
make
targets seamlessly handle the container orchestration.
If, for whatever reason, you must opt-out of executing development tasks within
containers, set the SKIP_DOCKER
environment variable to true
, but be aware
that by doing so, the success or failure of development-related tasks, tests,
etc. will be dependent on the state of your system, with no guarantee of the
same results in CI. For running the full array of targets in this mode, you’ll
need Go 1.16+ and JavaScript ecosystem utilities (yarn, npm, node,
etc.)
Developing on Windows
All development-related tasks should “just work” on Linux and Mac OS systems. When developing on Windows, the maintainers strongly recommend utilizing the Windows Subsystem for Linux 2. See more details here.
Working with Go Code
To run lint checks:
$ make lint-go
To run the unit tests:
$ make test-unit-go
Working with JS Code (for the Brigade Worker)
To lint the Javascript files:
$ make lint-js
To run the tests:
$ make test-unit-js
To clear the JS dependency cache:
$ make clean-js
Building Source
Note that we use kaniko-based images to build Brigade component Docker images. This is a convenient approach for running the same targets in CI where configuring access to a Docker daemon is either unwieldy or rife with security concerns. This also means that images built in this manner are not preserved in the local Docker image cache. For the image push targets discussed further down, the images are re-built with the same kaniko image, albeit with the corresponding publishing flags added.
To build all of the source, run:
$ make build
To build just the Docker images, run:
$ make build-images
To build images via Docker directly and preserve images in the Docker cache, run:
$ make hack-build-images
To build all of the supported client binaries (for Mac, Linux, and Windows on amd64), run:
$ make build-cli
To build only the client binary for your current environment, run:
$ make hack-build-cli
Pushing Images
By default, built images are named using the following scheme:
brigade-<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). Here we use the targets that utilize Docker directly, so that the
images exist in the local cache:
$ DOCKER_ORG=krancour make hack-build-images
To build for the krancour
org on a different registry, such as quay.io
:
$ DOCKER_REGISTRY=quay.io DOCKER_ORG=krancour make hack-build-images
Now you can push these images. Note also that you must be logged into the registry in question before attempting this.
$ make hack-push-images
Otherwise, when using the kaniko-based targets, the images will be built and pushed in one go. Be sure to export the same registry/org values as above:
$ DOCKER_REGISTRY=quay.io DOCKER_ORG=krancour make push-images
Minikube configuration
Start Minikube with the following required addons enabled:
- default-storageclass
- storage-provisioner
To view all Minikube addons:
$ minikube addons list
Additionally, for local development, it will be efficient to enable the
registry
addon to set up a local registry to push images to. See full
details in the registry addon docs. Here is an example on how to enable
the addon and redirect port 5000 on the Docker VM over to the Minikube machine:
$ minikube addons enable registry
$ docker run -d --rm --network=host alpine ash \
-c "apk add socat && socat TCP-LISTEN:5000,reuseaddr,fork TCP:$(minikube ip):5000"
Now to build and push images to the local registry and deploy Brigade, simply run:
$ export DOCKER_REGISTRY=localhost:5000
$ make hack
During active development, the overall flow might then look like this:
$ # make code changes, commit
$ make hack
$ # (repeat)
$ # push to fork and create pull request
For finer-grained control over installation, 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 charts/brigade > myvalues.yaml
$ open myvalues.yaml # Change all `registry:` and `tag:` fields as appropriate
From here, you can install or upgrade Brigade into Minikube using the Helm directly:
$ helm upgrade --install -n brigade brigade charts/brigade -f myvalues.yaml
To expose the apiserver port, run the following command:
$ make hack-expose-apiserver
The root user password will be auto-generated if not overridden in the values
file (via apiserver.rootUser.password
). To retrieve its value after install,
follow the steps provided in the Notes
section after deployment.
You can then log in to the apiserver with the following brig
command:
$ brig login -s https://localhost:7000 -r -p <root user password> -k
To create your first Brigade project, check out projects to see how it’s done.
Kind configuration
You can also use kind for your day-to-day Brigade development workflow. Kind has a great quickstart that can be found here.
The Brigade maintainers use kind heavily and there currently exists a helper
script that will create a new kind cluster with a local private registry
enabled. It also configures nfs as the local storage provisioner and maps
localhost:31600
to the API server port. To use this script, run:
$ make hack-new-kind-cluster
Now you’re ready to build and push images to the local registry and deploy Brigade:
$ export DOCKER_REGISTRY=localhost:5000
$ make hack
Normally, the root user password will be auto-generated if not overridden in the
values file (via apiserver.rootUser.password
). For the hack
targets, this
value is hard-coded to F00Bar!!!
.
You can then log in to the apiserver with the following brig
command:
$ brig login -s https://localhost:31600 -r -p 'F00Bar!!!' -k
To create your first Brigade project, check out projects to see how it’s done.
When you’re done, if you’d like to clean up the kind cluster and registry resources, run the following commands:
$ kind delete cluster --name brigade
$ docker rm -f kind-registry
Running Brigade inside a remote Kubernetes cluster
Some developers use a remote Kubernetes cluster instead of Minikube or kind.
To run a development version of Brigade inside of a remote Kubernetes cluster, you will need to do two things:
- Make sure you push the Brigade Docker images to a registry the cluster can access
- Export the correct values for
DOCKER_REGISTRY
and/orDOCKER_ORG
prior to runningmake hack
Running the Integration Tests
Once you have Brigade running in a Kubernetes cluster, you should be able to run the integration tests, which will verifies basic Brigade functionality via the following checks:
- Logs into the apiserver
- Creates projects of varying types
- Creates an event for each project
- Asserts worker and job logs and statuses for each event
To run the tests, issue the following command:
$ make test-integration
See the tests directory to view all of the current integration tests.