Build ML data pipelines that scale to a cluster.
Recent revisions are kept. Restoring one creates a new revision.