We operate a website in AWS that uses Azure as the data source. Azure is populated by transactional replication from our customers. Once every 3 or 4 months a new version of the customer database is released. My company does not make this customer database. More often than not, the new version does not come with an update script and all the customer information is wiped and the new database version is freshly created.
How do I maintain an uninterrupted history of the customer data in the cloud through these updates? I see only two options: create a new database for each version, create a new schema for each version. The first will get expensive over time, the second feels hacky. If it's such a big deal should I just make my own update scripts?