I am not sure we can have a spawn wrapping the whole process lifecycle (and even if we do, it could last dozens of minutes?).
The idea is not to wrap the whole process lifecycle, but each “cycle”, from receiving the message to updating the state to sending the diff back.
For example, we could have a mount event that would be wrapping the whole (or a good part of) defhandle_info({Phoenix.Channel, auth_payload, from, phx_socket}, ref).
