Brigade 1.0: Event-Driven Scripting for Kubernetes

Introduction

On October 17, 2017, the Microsoft/Deis team released Brigade 1.0, an event-driven scripting engine that lets you author CI/CD-style workflows in JavaScript and execute them on Kubernetes. Instead of heavyweight pipeline DSLs, Brigade focuses on developer-friendly scripts triggered by webhooks from GitHub, Docker Hub, Azure Container Registry and arbitrary HTTP sources.

Brigade Building Blocks

• Gateways receive external events and translate them into Brigade payloads. The GitHub gateway ships out of the box, with others easily authored in Go or JavaScript.
• Brigade Controller watches for new events, reads a project’s brigade.js, and launches worker pods.
• Workers run Node.js to interpret brigade.js, orchestrating containerized jobs via the Kubernetes API.
• Jobs wrap container images, commands and concurrency options; steps can run sequentially or in parallel, enabling build/test/deploy flows.

Projects are defined with a brigade-project.yaml (Kubernetes Secret) storing repo references, secrets and service principals.

Sample Workflow

const { events, Job } = require("brigadier")

events.on("push", async event => {
  let test = new Job("jest-tests", "node:8")
  test.tasks = [
    "cd /src",
    "npm install",
    "npm test"
  ]

  let dockerBuild = new Job("publish-image", "docker:17.09")
  dockerBuild.privileged = true
  dockerBuild.tasks = [
    "dockerd-entrypoint.sh &",
    "sleep 3",
    "docker build -t example/webapp:$BRIGADE_COMMIT .",
    "docker push example/webapp:$BRIGADE_COMMIT"
  ]

  await test.run()
  await dockerBuild.run()
})

Each event spins up an isolated worker pod, mounts the repo (via brigade-git-sidecar), and executes the JavaScript-defined job graph.

Use Cases

• Lightweight CI/CD for teams wanting GitHub-driven pipelines without Jenkins or Spinnaker overhead.
• ChatOps integrations responding to slash commands that orchestrate Kubernetes jobs.
• Container-native automation such as nightly security scans, config generation or data hydration tasks.
• Prototype glue code combining services (e.g., trigger Helm releases after image pushes).

Operational Notes

• Brigade stores project secrets in Kubernetes; integrate with RBAC/NetworkPolicies to restrict access.
• Worker pods mount Docker-in-Docker for builds; consider Kaniko or brigade-k8s-job for rootless workflows.
• Logs stream through brigade CLI; long-running pipelines benefit from LogStream/Elastic integration.
• Scaling relies on Kubernetes; ensure the cluster has capacity for bursts of worker pods.

Ecosystem Momentum

The 1.0 release arrived alongside:

• Kashti dashboard (alpha) for visualizing projects and job runs.
• Early Azure Container Registry gateway for image-triggered workflows.
• Community scripts targeting Helm releases, Slack notifications and ChartMuseum promotions.

With CNCF Sandbox admission on the horizon, Brigade positioned itself as a flexible, developer-friendly bridge between GitOps tooling and traditional CI pipelines.

Summary