Is it a big deal to just switch to postgres/mariadb/mysql and use other open source technology add-ons or patterns to scale?
Aha, finally a question on this site that I can answer from a professional capacity.
Switching database for an application can range from trivial to very complex depending on the nature of query you run on it. For small applications where you might write to it infrequently and do simple select on the data a few times a second switching database would be a small project where you just have to come up with DDL for new schema, then do a data migration.
For larger applications with 3 digit number of users all constantly querying it at the same time, plus heavy lifting jobs like daily batch processing with more complicated reporting queries with lots of joins then it could be a year long project. Performance for large queries can be very sensitive to caching changes because they affect the query optimizer logic and even small changes can result in queries switching to a new query plan with disastrous consequences for execution time. I've seen at work when a normal 1 minute query in production picking up a bad query plan in non-production because the developer ran a gather stats on tables that is out of ordinary to normal production weekly usage pattern (he just happened to gather stats on table when the table was empty) and the query run time balloon to several hours instead.
Switching entire database means every query that has heavy dependency on correct query plan needs to be rechecked and performance tested, as different DBS have entirely different query optimizers logic.
When we switched from Oracle 13 running on locally hosted machine in our data center to Oracle 19 running on AWS RDS the project took more than a year with lots of headache involved in performance testing and optimizing queries in the middle, and that's just from one version of Oracle to another.