Missing documentation about container registry

Regarding Kubeflow [1], guide [2] employs [3], however, there is no documentation for [3].
This should be better documented in GCP documentation, namely:

- A documentation page for using [3].
- It should explain what is included within [3] (packages, dependencies etc, ...).
- How does [3] fit into the broader framework of Kubeflow pipelines.
- [2] seems outdated , there should be a documented version of [2] which uses the updated versions of beam and python.

--------------------
[1] https://github.com/GoogleCloudPlatform/training-data-analyst/blob/master/courses/machine_learning/deepdive/06_structured/pipelines/README.md
[2] https://towardsdatascience.com/how-to-create-and-deploy-a-kubeflow-machine-learning-pipeline-part-1-efea7a4b650f
[3] gcr.io/ml-pipeline/ml-pipeline-dataflow-tft