This counts frames whose sequence number was higher than expected, meaning one or more frames in between were lost (a discontinuity, or 'gap', in the sequence rather than a simple reorder). An admin uses this alongside the OOO counter to distinguish loss-with-gap events from pure reordering, since both can occur in the same test run but point to different root causes (congestion drops vs. multipath/LAG reordering). For example, a rising cienaCesBenchmarkTestInstanceStatsRxDiscSeqNumPkts count paired with a flat OOO counter tells the engineer the problem is genuine frame loss creating sequence gaps, not reordering.