Possibly! We didn't consider databases because this logic is implemented in agents injected in our users' applications. On top of limiting what we had access to, it also put a lot of pressure on us to minimise our CPU & memory overhead.
Because of that, we believed using a database (even in-memory) would have resulted in an unacceptable overhead.
Because of that, we believed using a database (even in-memory) would have resulted in an unacceptable overhead.