The approach always depends on which engine causes workqueue to build. At times this might be shown under "workqueue status" output, but there may be situations when this info is not visible.
TAC has ways to gather more info about engines that may be causing issues via tunnel access. However once there is an identification, message filters can be used to drop emails or skip engines based on required conditions to clear up the workqueue.
It wont be effective in every situation, but it surely can help. If you increase the maximum concurrent connections, it gives room to senders to inject more emails via parallel connections (which ofcourse depends on mail flow policy settings too). It's hard to say, but it can influence the workqueue build up if too many connections are accepted as result of the change.
