After a couple of check-ins from the sensors the message will go away
Additional Information
Kernel Queue is hard coded to 12k for all OS's. By considering physical memory size on the box this value can be higher, resulting in the kernel not dropping events.
At start, the kernel queue caches events before the user level service can start
Excessive event loss can also be seen
If event types have been disabled in the sensor group
If the sensor service is not running at all. Run sc query carbonblack in cmd to confirm