CI/CD System


Difficulty	Advanced
Team Size	3-5 people
Time	~35-45 hours
Demo-ready by	Step 5
Prerequisites	Node.js, Docker, Git, shell scripting
Built by	GitHub Actions, Jenkins, GitLab CI, CircleCI

Skills you'll earn: Pipeline DSL, container-based builds, parallel test execution, artifact management, deployment gates

Start by running a shell script on git push. End with a pipeline system that builds in containers, runs tests in parallel, manages artifacts, and gates deployments.

Step 1: Run a script on git push (~2-3 hours)

The core of CI: something happens automatically when code changes.

Set up a Git post-receive hook on a bare repo
The hook runs a shell script: pull the latest code, run npm test (or equivalent)
Print pass/fail to a log file
Trigger this by pushing to the repo

You now have: Automated test execution on push.

Step 2: Build a CI server with an API (~3-4 hours)

Shell scripts in git hooks don't scale.

Build a server that accepts webhook payloads from GitHub or Gitea
On receiving a push event, clone the repo and run the pipeline
Store build results: status, logs, duration, commit hash
API: GET /builds to list recent builds, GET /builds/:id for details

You now have: A CI server.

Step 3: Run builds in containers (~4-5 hours)

Building on the host is dangerous. A malicious package.json script owns your server.

Run each build inside a Docker container
Mount the repo into the container, run the build command, capture stdout/stderr
Destroy the container after the build completes
Use different base images per language (node, python, golang)

You now have: Isolated, reproducible builds.

Step 4: Pipeline definitions (~3-4 hours)

Real projects need more than "run tests." They need build → test → lint → deploy as a sequence.

Define pipelines in a YAML file (.ci.yml) in the repo root
Parse the YAML into stages, each with a name, image, and commands
Execute stages sequentially, stop on first failure
Report per-stage status and timing

You now have: Multi-stage pipelines.

Step 5: Parallel execution (~4-5 hours)

Linting doesn't depend on tests. Run them at the same time.

Support parallel steps within a stage
Spawn multiple containers simultaneously
Aggregate results: the stage passes only if all parallel steps pass
Show a pipeline graph with parallel lanes

You now have: Parallel CI pipelines.

Step 6: Artifacts and caching (~3-4 hours)

Save build outputs (binaries, Docker images, test reports) as artifacts
Artifact storage with retention policies (keep last 10 builds)
Cache dependencies between builds (node_modules, pip cache) to speed up subsequent runs
Mount cache volumes into build containers

Step 7: Deploy gates and environments (~3-4 hours)

Add a deploy stage that pushes to staging or production
Manual approval gates: pipeline pauses until someone approves
Environment locks: only one deploy to production at a time
Rollback: re-run a previous successful deploy pipeline

Useful Resources

Where to go from here

Matrix builds (test across multiple language versions and OS)
Secret injection into build containers (from your config/secrets system)
Build status badges for README files
Scheduled pipelines (nightly builds)
Pipeline-as-code with reusable templates

CI/CD System ​

Step 1: Run a script on git push (~2-3 hours) ​

Step 2: Build a CI server with an API (~3-4 hours) ​

Step 3: Run builds in containers (~4-5 hours) ​

Step 4: Pipeline definitions (~3-4 hours) ​

Step 5: Parallel execution (~4-5 hours) ​

Step 6: Artifacts and caching (~3-4 hours) ​

Step 7: Deploy gates and environments (~3-4 hours) ​

Useful Resources ​

Where to go from here ​