Labels

In order to make it easier to identify objects that are all part of the same conceptual pipeline, custom labels set on resources used by Tekton Pipelines are propagated from more general to more specific resources, and a few labels are automatically added to make it easier to identify relationships between those resources.

Propagation Details

For Pipelines executed using a PipelineRun, labels are propagated automatically from Pipelines to PipelineRuns to TaskRuns and then to Pods. Additionally, labels from the Tasks referenced by TaskRuns are propagated to the corresponding TaskRuns and then to Pods.

For TaskRuns executed directly, not as part of a Pipeline, labels are propagated from the referenced Task (if one exists, see the Specifying a Task section of the TaskRun documentation) to the corresponding TaskRun and then to the Pod.

For Conditions, labels are propagated automatically to the corresponding TaskRuns and then to Pods.

Automatically Added Labels

The following labels are added to resources automatically:

  • tekton.dev/pipeline is added to PipelineRuns (and propagated to TaskRuns and Pods), and contains the name of the Pipeline that the PipelineRun references.
  • tekton.dev/pipelineRun is added to TaskRuns (and propagated to TaskRuns and Pods) that are created automatically during the execution of a PipelineRun, and contains the name of the PipelineRun that triggered the creation of the TaskRun.
  • tekton.dev/task is added to TaskRuns (and propagated to Pods) that reference an existing Task (see the Specifying a Task section of the TaskRun documentation), and contains the name of the Task that the TaskRun references.
  • tekton.dev/clusterTask is added to TaskRuns (and propagated to Pods) that reference an existing ClusterTaskand contains the name of the ClusterTask that the TaskRun references. For backwards compatibility, TaskRuns that reference a ClusterTask will also receive tekton.dev/task.
  • tekton.dev/taskRun is added to Pods, and contains the name of the TaskRun that created the Pod.
  • app.kubernetes.io/instance and app.kubernetes.io/component is added to Affinity Assistant StatefulSets and Pods. These are used for Pod Affinity for TaskRuns.

Examples

Finding Pods for a Specific PipelineRun

To find all Pods created by a PipelineRun named test-pipelinerun, you could use the following command:

kubectl get pods --all-namespaces -l tekton.dev/pipelineRun=test-pipelinerun

Finding TaskRuns for a Specific Task

To find all TaskRuns that reference a Task named test-task, you could use the following command:

kubectl get taskruns --all-namespaces -l tekton.dev/task=test-task

Finding TaskRuns for a Specific ClusterTask

To find all TaskRuns that reference a ClusterTask named test-clustertask, you could use the following command:

kubectl get taskruns --all-namespaces -l tekton.dev/clusterTask=test-clustertask