Part of the value of doing an ETL pipeline via streaming replication is you get the full history of data in a table. An SCD type 2 table where each row also has a valid_from and valid_to timestamp column.
How would someone do the same thing with this architecture?