In case of fleet-wide software updates (which e.g. observable by an upgrade of debmonitor-client this morning), load on the single CPU debmonitor instances can spike to 100%, let's add a second CPU to these hosts for some more head-room.