pipeline
A cloud-native Pipeline resource.
An open-source system for defining build, test, and deployment workflows inside Kubernetes using configuration files, where each step runs in its own container and tasks can be reused across pipelines.
Tekton Pipelines is an open-source system for defining and running automated software delivery workflows, such as building code, running tests, and deploying applications. It runs inside Kubernetes, which is a platform used to manage and run software in containers across multiple servers. If you are not already using Kubernetes in your infrastructure, Tekton is not something you would adopt standalone.
The core idea is that you describe your build and deployment steps as structured configuration files, similar to how you describe other things in Kubernetes. A pipeline is composed of tasks, and each task runs inside its own container. Because the tasks are independent, they can be reused across multiple pipelines, and the same pipeline can deploy to different environments without modification.
Tekton is designed to be flexible about what tools you use at each step. For example, the step that builds a container image can use different underlying tools interchangeably, as long as they fulfill the expected contract. This keeps pipelines from being tightly locked to one particular build tool.
The project is maintained under the Continuous Delivery Foundation, a vendor-neutral open-source foundation, and contributions follow standard open-source practices with a guide for new contributors and a defined development setup process. The README is fairly brief and points readers to external documentation for installation and tutorials.
This repository is primarily relevant to software engineering teams that already operate Kubernetes clusters and want a Kubernetes-native way to define their continuous integration and continuous delivery workflows, rather than relying on external CI/CD services.
Where it fits
- Define reusable build and test steps as Kubernetes configuration files instead of scripting them inside a third-party CI service.
- Run the same deployment pipeline against staging and production environments without duplicating configuration.
- Replace an external CI/CD service with a self-hosted, Kubernetes-native pipeline system your team controls entirely.