This is the timestamp of when the node's throughput, measured in packets per second, hit its recorded maximum since the last reset. Unlike the character-rate peak time tracked separately, this one is the relevant metric for packet-switched traffic (such as X.25 or frame relay), so an admin uses it to correlate a packet-rate spike with a specific batch job or network event. For example, a peak coinciding with a scheduled data-replication run would tie the spike to that job rather than to a fault condition.