OCP on Cloud parquet files are currently created during the download flow, this processing should be split into a separate celery task that runs Trino SQL. Large impact to the data processing pipeline, no impact to customer facing API/UI, update and document existing masu endpoint to use new flow.

This should be done one provider at a time in order to reduce risk.

Design Doc

This will be released one provider at a time to reduce risk of a larger change. AWS will be the first.  Targeting stage by 2024/08/15.