does typesense have any rate limit mechanism for self hosted? is there any recommended implementation for that or are we expected to create a front facing api to middleman requests to typesense?