I've decided on a slightly different approach for my application and would like
feedback:
``Each client connected to the server has its own queue of requests for data
independent of other clients, when the client makes a request for data that is
not cached and needs to be fetched from the database, this request is added to
the list and a future is returned that will complete when the request is
fulfilled. Every 400ms the client gets the first item from the request list and
executes the database query and completes the future with the database response
(also adds the returned data to the cache), returning control to the application
to respond in a timely manner. I did this to avoid massive requests to the
database, limiting each client's requests by time.``
This is good or bad? Do you think I should just use the database directly
without a "rate limit"?

PS: i'm using allographer with PostgreSQL in Nim.