Tekton Pipelines environment and resources

The following information and resources are available by default to an IBM Cloud® Continuous Delivery Tekton PipelineRun.

`PipelineRun` annotations

Table 1 describes the annotations that are included in a Tekton PipelineRun.

Table 1. PipelineRun annotations
Run Annotation	Description
`devops.cloud.ibm.com/build-number`	The build number for the pipeline. This number is the cumulative total of pipeline runs from all triggers.
`devops.cloud.ibm.com/listener`	The Tekton `eventlistener` that is mapped to the trigger that initiated this pipeline run.
`devops.cloud.ibm.com/pipeline-id`	The unique ID of the pipeline that is used for this run.
`devops.cloud.ibm.com/trigger-name`	The name of the trigger that initiated this pipeline run.
`devops.cloud.ibm.com/trigger-type`	The type of the trigger that initiated this pipeline run. The type can be either `manual`, `timer`, `scm`, or `generic`.
`devops.cloud.ibm.com/triggered-by`	The user who initiated the trigger. This value is the IBMId of either the user who pressed run for `manual` triggers, the user who last saved the trigger for `timer` triggers, or the user who performed the Git action for `scm` triggers. This value is empty for `generic` triggers.
`devops.cloud.ibm.com/pipeline-run-url`	The URL of the details page for this pipeline run.
`devops.cloud.ibm.com/tekton-pipeline`	The unique ID of this pipeline run.

Accessing annotations

You can access annotations from the tasks within your pipeline definition YAML file. The following example shows how to access the devops.cloud.ibm.com/build-number annotation. To provide that annotation as an environment variable, replace it with one of the run annotations from Table 1. PipelineRun annotations.

apiVersion: tekton.dev/v1beta1
kind: Task
metadata:
  name: simple-task
spec:
  steps:
    - name: simple-step
      image: ubuntu
      env:
        - name: BUILD_NUMBER
          valueFrom:
            fieldRef:
              fieldPath: metadata.annotations['devops.cloud.ibm.com/build-number']
      command: ["bash", "-c"]
      args:
        - echo $BUILD_NUMBER;
          echo "COMPLETED"

PipelineRun `ConfigMap` and `Secret`

The Continuous Delivery Tekton PipelineRun resource creates a specific ConfigMap and Secret for environment properties. Secure properties are available in the secure-properties Kubernetes Secret, including any Tool integration properties for which the selected field contains a secure value. Nonsecure properties are available in the environment-properties Kubernetes ConfigMap. The keys are the name of the field that is provided in the pipeline authoring user interface.

When you access a ConfigMap or Secret, make sure that you locate the correct object name and references so that your pipeline can successfully complete.

Accessing individual values

You can access individual key-values within a Task. The following code snippet uses the sample apikey property name (with secure properties) and the environment property name (with text properties). These property names are set in the delivery pipeline Properties page.

apiVersion: tekton.dev/v1beta1
kind: Task
metadata:
  name: cm-echo-props
spec:
  steps:
    - name: cm-show-props
      image: ubuntu
      env:
        - name: SECURE_VALUE
          valueFrom:
            secretKeyRef:
              name: secure-properties
              key: apikey
        - name: ENVIRONMENT
          valueFrom:
            configMapKeyRef:
              name: environment-properties
              key: environment  
      command: ["/bin/bash", "-c"]
      args:
        - echo -e "environment from ConfigMap is >>";
          echo $ENVIRONMENT;
          echo "";
          echo -e "apikey from Secrets is >>";
          echo $SECURE_VALUE

Accessing all values

You can add all of the key-value pairs from the ConfigMap and Secret to your Task environment:

apiVersion: tekton.dev/v1beta1
kind: Task
metadata:
  name: cm-secrets-props
spec:
  steps:
    - name: cm-show-full-env
      image: ubuntu
      envFrom:
        - configMapRef:
            name: environment-properties
        - secretRef:
            name: secure-properties     
      command: ["/bin/bash", "-c"]
      args:
        - echo -e "The environment for this Step is ";
          env

Managed worker virtual machine (VM) sizing

When you run a pipeline by using the IBM Managed Worker pool, a VM with a specific default memory is allocated. Although most jobs can run successfully with the provided memory, certain pipelines require extra memory for intensive tasks.

Users can specify a label on their tasks to indicate whether a task requires more (or less) memory for a specific task. This ability to identify the specific amount of resources that are required benefits resource usage and eventual cost savings.

To indicate which VM profile to apply to a specific task within a Tekton pipeline, add the runtimeClassName label to the task with one of the following VM values. If no label is provided, the default VM profile is used.

kata-tiny: 128Mi
kata-small: 2Gi
kata-medium: 4Gi (default)
kata-large: 8Gi

For example:

apiVersion: tekton.dev/v1beta1
kind: Task
metadata:
  name: task1
  labels:
    runtimeClassName: kata-tiny
    and so on.

You can also specify the same configuration in the PipelineRun as part of the TriggerTemplate. By specifying the runtimeClassName in the PipelineRun, users can select the VM profile that they want to use without changing the task definitions.

For example:

apiVersion: tekton.dev/v1
kind: PipelineRun
metadata:
  generateName: pipeline-run-
spec:
  pipelineRef:
    name: pipeline
  taskRunSpecs:
    - pipelineTaskName: task1
      taskPodTemplate:
        runtimeClassName: kata-tiny

Learn more about Tekton delivery pipelines

To learn more about Tekton and Continuous Delivery, see Tekton: A Modern Approach to Continuous Delivery.