some indexer users like the Grants Data Portal, need all the data at once, it will be better for us and them if we give them all the data in a suitable format, as the GraphQL will be really insufficient and use up a lot of resources.
I would suggest to implement a daily parquet export of all tables to s3.