Tekton Fundamentals

Tekton turns software delivery work into Kubernetes resources. You describe the work in YAML. Tekton controllers then create Pods, track execution, and report status through the Kubernetes API.

This approach matters when your delivery platform already centers on Kubernetes. Your pipelines use the same API patterns, namespaces, service accounts, and reconciliation model as other cluster workloads. Tekton gives you building blocks rather than a complete hosted CI service.

The mental model

Separate definitions from executions:

• A Step runs one container image and performs one operation.
• A Task defines an ordered set of Steps.
• A TaskRun executes one Task.
• A Pipeline defines a graph of Tasks.
• A PipelineRun executes one Pipeline.

A Task runs as one Kubernetes Pod. Its Steps run in order inside that Pod. A Pipeline can run independent Tasks in parallel, while dependencies form a directed acyclic graph. Each Task in that Pipeline gets its own Pod.