Since our database is not updated often, we should probably just trigger the backup to take place upon every write (i.e. when a new dataset is uploaded).
Consequently, this would allow us to also implement a faster and more stable rollback mechanism, as the currently envisioned rollback mechanism would take a stored, raw .xlsx file and run the database initialization script on it.
Description
We should have database backups.
Since our database is not updated often, we should probably just trigger the backup to take place upon every write (i.e. when a new dataset is uploaded).
Consequently, this would allow us to also implement a faster and more stable rollback mechanism, as the currently envisioned rollback mechanism would take a stored, raw .xlsx file and run the database initialization script on it.