We should add a counter of how many tasks which were marked as non-deprioritizable (either via the internal marker or the client based exemption) to the server status output. This will give us an indicator if too many requests are being exempted from deprioritization in a cluster.