With S3 sync we can potentially keep multiple local checkpoints around. We need an API to list those checkpoints and their disk footprint in order to understand their impact on the total storage footprint of the pipeline.