Skip to content
go-tektasker

Getting Started

Introduction

In this guide, we will initiate a go project and turn it into an usable Tekton Task using go-tektasker.

Requirements

Ensure you followed the installation instructions.

Steps

Step 1: Initiate the project

# Create a working directory for your task
mkdir my-task
cd my-task

# Initiate a Go module
go mod init example.com/my-task

# Scaffold an example task with an opinionated development environment
go-tektasker init my-task

Step 2: Get familiar with the structure

You will find a task.go containing an example of a task using tektasker.

The opinionated Development Environment is also present, you can remove the .env and Taskfile.yaml if you prefer to use go-tektasker as a standalone.

This environment behavior defaults to:

  • Write your task manifest into the deployment/ directory.
  • Generate go helper code package in internal/tekton.
  • Build and push your task container image to ko.local (controlled by KO_DOCKER_REPO)
  • Apply your task to the currently used Kubernetes Context (controlled by APPLY_CONTEXT)

Step 3: Understand the markers

Tektasker uses a concept borrowed from kubebuilder: markers.

Markers are go comments with a special syntax that are used to mark your code constructs (structures, package, …). You can use markers to tell to Tektasker in which structure you would like to receive a specific parameter or result.

At the package level, the marker +tektasker:task is used to give a name to your task and specify its version and is mandatory.

You can see the list of available markers by using go-tektasker markers. If you want to get a more verbose help about a specific marker, specify it as an argument to the previous command.

Step 4: Follow the instruction in task.go

You will be asked to generate the go helper code and remove the comments before the MustRead and MustWrite instructions.

task generate

Feel free to change the format of the output message that will be stored in the lm result.

By now, it should be clear to you that:

  • MustRead tries to read a Tekton Task parameter or panic if it fails.
  • MustWrite tries to write your result back to Tekton or panic if it fails.

If you prefer to not panic and handle errors gracefully, those functions exists without the Must prefix.

Step 5: Create the YAML manifest

task manifest

This will create a base Kustomization for your task in the deployment/ directory.

Step 6: Deploy your task

Now that you have your YAML generated, you can deploy your task.

If you use the Development Environment, just run task apply.

Otherwise, you need to manually:

  • Build and push a container image of your go project.
  • Create an overlay of the generated base Kustomization to replace the image of the step.
  • Apply the overlay you created on your cluster.

Conclusion

This was a short introduction to Tektasker. The project is still in alpha stage so the documentation and features should get richer in the future. Do not hesitate to give us feedback by: