I'm learning fastAPI - and encountered concept of concurrency. From point of view of database transactions assume that underlying driver is not 'async enabled', so to speak. For my API,for each call to it, it does a simple GET request to underlying database.If the bottleneck is actually the transaction times themselves, does this mean that:
- There is not much benefit in using async/await - as underlying driver doesn't support concurrency
- using a connection pool would be an alternative(I'm aware it's a different thing altogether)
- went to docs - learnt about await - but mostly they use psychopg(async)
- I expect connection pooling will provide a small advantage