What Are CI and CD?

Continuous Integration (CI) is the practice of frequently merging code changes into a shared repository and running automated tests to catch issues early. It helps development teams collaborate smoothly and catch bugs before they compound.

Continuous Deployment (CD) takes this a step further, it automatically delivers tested code to production (or staging), enabling businesses to ship features faster, gather real user feedback, and iterate quickly.

Both practices rely on a handful of engineering fundamentals: Test-Driven Development (TDD), sensible branching models, well-defined workflows, and solid version control (VCS) habits.

Continuous Integration in Action

Without diving into any specific branching model, the diagram below illustrates a typical CI workflow:

Here's what happens at each step:

  1. Parallel development: Developer A and Developer B work on separate branches simultaneously.
  2. Early merging: Once a unit of work is complete (e.g., a function or class method), it's merged into the main branch.
  3. Automated testing: Tests run automatically. If any test fails, the team is immediately notified.
  4. Sync & share: The main branch is merged back into each developer's branch, keeping everyone's work in sync.

Continuous Deployment in Action

Once code passes CI, CD takes over. The diagram below shows a typical deployment pipeline:

The flow is straightforward:

  1. Code is merged and tested, all CI checks pass.
  2. Automated build triggers, a new artifact is generated.
  3. Automatic deployment, the code is pushed to Staging or Production servers without manual intervention.

A Note on Code Reviews

In many real-world scenarios, code reviews are a required step before merging. Common strategies include pull requests and review requests. While important, they're outside the scope of this tutorial.

Tools of the Trade

To make all of this work, the following tools are commonly used:

  • Jenkins, defines and orchestrates test and build tasks.
  • Capistrano, deploys code across multiple servers using Git.
  • Git, the version control backbone.

Hook scripts can be configured to trigger these tasks automatically, but this tutorial focuses on the manual process.