Concourse — Container-Native CI/CD with Pipelines as Code

What it is

Concourse is an open-source CI/CD system where every task runs in its own OCI container. Pipelines are defined in declarative YAML, and the resource model makes dependencies between jobs explicit and reproducible. Unlike Jenkins or GitHub Actions, Concourse treats pipelines as first-class code with no implicit state.

Concourse is for DevOps teams and platform engineers who need reproducible, auditable CI/CD pipelines. Its container-per-task model ensures that builds run identically regardless of where the worker runs.

How it saves time or tokens

Concourse's resource model eliminates hidden dependencies. Every input and output is declared explicitly, so you never encounter 'works on my machine' pipeline failures. When a pipeline breaks, you know exactly which resource changed.

The container-per-task isolation means no shared state between jobs. You do not need to clean workspaces, manage tool versions on shared runners, or worry about one job corrupting another's environment.

How to use

1. Start Concourse with Docker Compose:

curl -O https://concourse-ci.org/docker-compose.yml
docker-compose up -d

1. Log in with the fly CLI:

fly -t local login -c http://localhost:8080 -u test -p test

1. Define a pipeline in YAML and set it:

# pipeline.yml
resources:
- name: repo
  type: git
  source:
    uri: https://github.com/org/app.git
    branch: main

jobs:
- name: test
  plan:
  - get: repo
    trigger: true
  - task: run-tests
    config:
      platform: linux
      image_resource:
        type: registry-image
        source: { repository: node, tag: '20' }
      inputs:
      - name: repo
      run:
        path: sh
        args: ['-c', 'cd repo && npm ci && npm test']

fly -t local set-pipeline -p my-pipeline -c pipeline.yml
fly -t local unpause-pipeline -p my-pipeline

Example

A multi-stage pipeline with build, test, and deploy:

jobs:
- name: build
  plan:
  - get: repo
    trigger: true
  - task: compile
    config:
      platform: linux
      image_resource:
        type: registry-image
        source: { repository: golang, tag: '1.22' }
      inputs: [{ name: repo }]
      outputs: [{ name: binary }]
      run:
        path: sh
        args: ['-c', 'cd repo && go build -o ../binary/app .']
  - put: artifact-bucket
    params: { file: binary/app }

Related on TokRepo

• DevOps AI tools -- CI/CD and infrastructure tools
• Automation tools -- pipeline and workflow automation

Common pitfalls

• Concourse has a steeper learning curve than GitHub Actions or GitLab CI. The resource model is powerful but requires understanding inputs, outputs, and triggers before writing your first pipeline.
• Worker scaling is manual. Unlike cloud CI services, you manage your own Concourse workers. Under-provisioned workers cause queued builds.
• Concourse does not have built-in secret management. Integrate with Vault, AWS Secrets Manager, or CredHub for secure credential handling.