This 16-bit counter tracks how many times, since the last node or statistics reset, incoming data on this port was lost specifically because the node's CPU could not process it fast enough, as opposed to a line-quality problem. This distinguishes a CPU/processing bottleneck from a physical-layer fault, so a rising overrun count while CRC and framing errors stay flat points an admin toward CPU load or buffer starvation rather than a bad cable. For example, overrun errors climbing on a port at the same time cdx6500NMNPSstatscpuUtil shows sustained high values would confirm CPU saturation as the cause of the data loss.