Upload a YAML file to create a pipeline. The first file you upload becomes the pipeline’s default version.
Before you begin
Section titled “Before you begin”You need a role with create permission on the pipeline service and a .yaml file
written in Argo Workflows format or compiled from the Kubeflow Pipelines SDK. Use steps that each run
one container: a file that uses containerSet is refused. For a starting file, select See Example on
the create page, which shows a Hello World pipeline in both formats, or copy the one in
Get started with Pipelines.
Create the pipeline
Section titled “Create the pipeline”- In the side menu, under PRODUCTS, select Pipeline, then Create Pipeline.
- Leave Create a new pipeline selected.
- In Pipeline Name, keep the generated name or type your own, up to 30 characters.
- (Optional) In Description, add up to 300 characters.
- In Upload Pipeline File, add one
.yamlfile. - Select Create Pipeline.
The pipeline is listed in Manage Pipelines. Its Pipeline Type is Argo for an Argo Workflows file, Kubeflow for a Kubeflow Pipelines file, and None otherwise. When you create the first pipeline in a project, TIR also sets up an object storage bucket and an access key for the project.
For help with a file that is refused, see Troubleshooting Pipelines.
To run the pipeline, create a run. When the file changes later, create a version instead of replacing the pipeline.