When issuing a synchronous "systemctl daemon-reload" requests this will
send out a tonload of UnitRemoved+UnitNew signals, that we will queue
locally. When we wait for the message reply, we should be able to queue
those, hence bump the limits on locally queuing messages.