1. Last time of the sync is not to be confused with the time up to which data is loaded. Account for replication lag at the source.
2. The sync schedule storage - where will it be maintained?
3. Assuming the structure comparison will be done before every load - with an option to configure an approval for new field addition
4. Sync modes / 2 / deduped history - destination doesn’t support primary keys as far I understand. Unless you have staged the data and using it to join. I don’t understand. Will need you to explain what are we doing here.
5. CDC can’t capture deleted records — a very big caveat. Correct me if I’m wrong.