They run on Tekton, the open-source framework built for building CI/CD systems in Kubernetes. Each pipeline is a series of tasks. Tasks define steps. Steps run in lightweight containers. Everything runs inside the cluster. There is no hidden build farm. No fragile runner on a random VM.
With Openshift Pipelines, you define workflow as code in YAML. Declarative pipelines mean you can version them, audit them, and roll them back. Every change is tracked. Configuration lives alongside application code. Your builds become predictable and repeatable across environments.
Integrations with Kubernetes and OpenShift are baked in. Pipelines scale with your clusters. They use persistent volume claims if you need state, or run stateless for maximum speed. Secrets, service accounts, and namespaces give you fine control over permissions and isolation.
Pipeline resources describe what the pipeline consumes and produces — source code repos, container images, or artifacts in storage. This keeps pipelines portable and self-contained. Tasks can be reused across projects. Reusability increases velocity without cutting discipline.