Tekton Pipelines

What it is

Tekton Pipelines is a Kubernetes-native framework for building CI/CD systems. It defines pipelines, tasks, and triggers as Kubernetes Custom Resources (CRDs) — no separate orchestration server, no proprietary DSL runtime. Everything runs as pods on your cluster. It started as the CI/CD portion of Knative before being split out as its own CNCF project. It's the plumbing layer, not a finished CI/CD product — most teams either build custom pipelines on top of it or use it as the engine underneath something like Tekton Dashboard, Jenkins X, or OpenShift Pipelines (Red Hat's productized distribution).

Who builds it and why

Primarily driven by Red Hat (which ships it as the backbone of OpenShift Pipelines) alongside Google, IBM, and a long list of enterprise contributors who need CI/CD that runs natively on Kubernetes without a separate control plane. 435 contributors is a healthy number for a project this scoped — it indicates real cross-vendor investment, not a single-company side project. The motivation is straightforward: teams running everything on K8s don't want a bolted-on CI server with its own scheduler, storage, and RBAC model to reconcile against cluster-native tooling.

Production readiness signal

• 8,999 GitHub stars — modest for a project this foundational, but star count is a weak signal for infrastructure plumbing that's usually consumed indirectly (via OpenShift Pipelines, Tekton Dashboard, or custom wrappers).
• Latest release v1.14.0 with commits as recent as 2026-07-01 — active, not abandoned.
• 435 contributors — broad enough to survive any single company pulling back.
• CNCF maturity level: not publicly available (confirm current graduation status directly with CNCF's landscape before treating this as a factor in vendor risk assessment).
• Apache-2.0, written in Go — standard, no licensing surprises.

Bottom line: this is mature, boring-in-the-good-sense infrastructure. The API has been stable since v1beta1/v1 GA. If it's breaking on you, it's almost always a config or resource-request problem, not the project.

Who should use this

• Platform teams building an internal developer platform on Kubernetes who want CI/CD as CRDs, controllable via GitOps and kubectl.
• Organizations already running OpenShift, where Tekton is the default pipeline engine.
• Teams that need pipeline steps to run as isolated pods with fine-grained resource limits, node affinity, and cluster-native secrets/RBAC.
• Anyone building a custom CI/CD product or internal tool who wants a battle-tested execution engine instead of writing one.

Who should NOT use this

• Small teams wanting CI/CD in an afternoon — Tekton has real setup overhead (CRDs, controllers, RBAC, storage for workspaces). GitHub Actions or GitLab CI will get you running in minutes; Tekton won't.
• Teams without a Kubernetes cluster already in production — standing up a cluster just to run Tekton is backwards.
• Anyone wanting a polished UI and dashboard experience out of the box — the base project is API/CRD-first; Tekton Dashboard exists but is thin, and you'll likely end up building your own UI or buying OpenShift Pipelines for the packaged experience.
• Teams needing mature, built-in artifact caching, matrix builds, or a large marketplace of pre-built actions — you'll be writing more of that yourself compared to GitHub Actions or CircleCI.

Alternatives

• GitHub Actions — far lower setup cost, huge ecosystem of pre-built actions, but you're tied to GitHub's execution model and hosted/self-hosted runner limits.
• Argo Workflows — same CRD-native philosophy on Kubernetes, arguably better suited to general DAG-based data/ML workflows in addition to CI/CD.
• Jenkins — more mature plugin ecosystem and UI, but carries legacy operational overhead and isn't Kubernetes-native by design (though it can run on K8s).

Pricing

Fully open source, Apache-2.0, no paid tier for the core project. Red Hat sells support and a packaged distribution (OpenShift Pipelines), but Tekton Pipelines itself has no license cost.