How Drone CI Transforms Scalable Continuous Delivery for Modern Software Teams

Creating a successful software company requires delivering valuable software quickly. To ensure high‑speed service we rely on a robust Continuous Delivery (CD) process supported by a solid Continuous Integration (CI) mechanism.

We use Git as the sole source‑code manager, with each project having its own repository. Development branches are used for temporary builds, while the main branch builds production releases.

The question then becomes how to manage Git operations (pull requests, merges) and deploy code in a controlled manner across environments. The answer is a CI/CD tool.

Initially we used GitHub Actions for CI and Ansible/AWX for CD. As our roadmap grew in component count and dependencies, this setup became limited. We needed a solution that could handle multiple libraries and packages efficiently.

Selection

After evaluating tools such as CircleCI, TravisCI, TeamCity, and Bamboo, we chose Drone CI (https://www.drone.io) as the core of our CI/CD pipeline.

Open‑source with a large community

Easy to install and maintain

Docker‑based; everything runs in containers

Native integration with GitHub, GitLab, Bitbucket, and many others

YAML‑based configuration following the “pipeline as code” principle

Scalable with auto‑scaling on major cloud providers

Rich ecosystem of community‑maintained plugins; custom plugins are easy to develop

Built‑in secret management (with optional external systems)

Drone Basics

Drone triggers on any Git repository action via webhooks. If a pipeline is defined (e.g., a .drone.yml file at the repo root), Drone parses it and executes the defined steps based on triggers such as branch name and event type.

Each step is described with a simple syntax, for example using the maven:3.6.3-jdk-11 image to run mvn clean and mvn install. The steps share a workspace (e.g., /drone/src) so artifacts from one step can be used in later steps.

Drone Services

Drone supports defining services (e.g., Redis, Postgres) that run alongside pipeline steps using Docker‑in‑Docker. The platform automatically starts the required service container and terminates it after the pipeline finishes.

Drone Plugins

A Drone plugin is simply a container that runs code. While Go is the preferred language, any language can be used. Custom plugins can be published to a registry and referenced in the pipeline.

Testing and Test Reports

We add test steps for unit, integration, Cypress, Postman, etc., each running in its own container. To collect and view test results we use the open‑source Allure Docker Service , which provides an API layer and a web UI for browsing reports.

Run tests in a container and write results to a shared filesystem

Use a custom Drone plugin to send the results to the Allure service via its API

For Cypress, one step runs the tests and stores results in /drone/src/cypress-results/allure, and a second step uploads them to the Allure service.

Artifacts

The final CI pipeline should produce artifacts usable in any environment (staging, pre‑prod, prod). Our platform stores Docker images in ECR and uses Nexus OSS to manage npm packages and Java libraries. Drone can create these artifacts and push them to the appropriate repositories.

In summary, we chose Drone as our CD engine because it integrates seamlessly with our existing tools, offers Docker‑native pipelines, supports services and plugins, and enables reliable artifact generation and test reporting for our engineering teams.