Currently the flow for ongoing requests allows itself to be interrupted by new requests at several points when only synchronous code is actually being executed. This allows high traffic to needlessly inflate response times for a large number of requests and makes the engine caching as implemented less effective.
Currently the flow for ongoing requests allows itself to be interrupted by new requests at several points when only synchronous code is actually being executed. This allows high traffic to needlessly inflate response times for a large number of requests and makes the engine caching as implemented less effective.