In the same situation described in QTBUG-99793 there is another problem.

The server runs smoothly for many hours but, at some point, its cpu usage falls down near to 0.1% whereas its memory usage constantly increases until the server crashes.

The server has a lot of remote objects periodically synchronized by corresponding timers in the same thread. This generates a lot of synchronization signals that insert a lot of events to handle in the event queue.
If the frequency by which events are queued is greater than the frequency by which events are processed, the queue will grow over time. I suppose this could be the cause of what happens.

Does some ways exists to monitor size and other information about event queues?