Consider using `boost::asio::thread_pool`
Currently, we're maintaining our own threadpool.
It might be worth considering using boost::asio::thread_pool instead.
I think it's a very good decision, But when will coroutines be supported in routing. It is very practical when the data comes from other server
There is also boost.mysql
The main branch already provides a coroutine interface for the HTTP portion of the client.
The feature/ws_coro branch is working on using coroutines for WS.
Using coroutines in the HTTP server handlers is still will probably follow next.
But none of this has anything to do with this particular issue (using a boost supplied threadpool instead of cooking our own).