How to use the Tekton plugin

Connect Kestra to Tekton Pipelines installed on Kubernetes and orchestrate PipelineRuns and TaskRuns from a flow.

Authentication

Tasks and triggers accept a connection object. Fabric8 can use the current kubeconfig or in-cluster service account automatically. For explicit authentication, provide masterUrl, oauthToken, and optionally caCertData. Store credentials in Kestra secrets and reference them with {{ secret('K8S_TOKEN') }}.

Tasks

PipelineRuns

Use pipelinerun.Create to create a run from pipelineRef or an inline spec, with params, workspaces, and serviceAccountName. Set wait: true to poll the Tekton Succeeded condition. A False condition fails the Kestra task; waitTimeout limits how long it waits and cancels the run if the timeout expires. Get, List, and Cancel read, filter, and cancel runs.

TaskRuns

Use taskrun.Create and taskrun.Get for single Tekton Tasks. Create supports taskRef, inline spec, parameters, workspace bindings, optional waiting, waitTimeout, and cancellation when the Kestra execution is killed.

Triggers

PipelineRun trigger

pipelinerun.Trigger polls terminal PipelineRuns selected by labelSelector. Each poll is bounded by limit (100 by default); use a focused label selector for busy namespaces. It stores the last emitted run name and completion time in the namespace KV Store, so repeated polls do not emit the same run twice. The trigger exposes the run name, namespace, status, and completion time to the execution.

Cluster requirements

The target cluster must expose the Tekton tekton.dev/v1 API. The plugin does not install Tekton or create Kubernetes namespaces.