Possible fix to #474, #453, #399. Connection Reset by Peer, ReadTimeOut:

• I noticed that connection reset by peer maybe could be solved by resetting the requests.session(). ReadTimeOut too maybe can be solved by resetting it. So, when a request is made and a exception is captured, the session is resetted and a retry is done. Retries limit was set to 5.

Made an alternate solution to #481:

• Made polling_thread an instance variable: since if Telebot is threaded, maybe is for the best to limit polling_thread instances to just one.
• The function stop_instance disables the Telebot instance (since it closes the worker_pool its threads can't be restarted), but it grants that the polling_thread is closed too, if some unhandled exception happened on _threaded_polling and the self.polling_thread wasn't closed.
• In WorkerThread's stop() function, added a self.continue_event.set(), just to make sure that the thread will get out of the while self._running: loop (could the thread be stuck on the self.continue_event.wait()?).