Distributing geographically generally comes up in large-scale systems. My point is that relational databases already support that. I know what cloud functions are. How do they not work with relational databases, any more than they don’t work with some NoSQL tool? Unless your cloud functions embed the database they’re using a remote service, right?
Because on peak traffic you could have thousands of cloud functions running, each with its own connection to the database. Typically SQL databases are much more strict on the number of concurrent connections as these consume more memory than on NoSql databases.
Edit:
IIRC I read that each connection to Postgres consumes 10MB or RAM.
Maybe the proposal above by itself isn't enough, but it's going to solve some of the many problems that put a restriction on max allowable connections in the first place. For example, I am not sure if there is anything in there to reduce the baseline per connection memory consumption, but maybe that will come up as the next problem to solve and will hopefully be solved sooner than later.
Also I’m not a big enterprise nor a bank.