Apinger reporting incorrect latency
If a gateway has an explicit monitor address, apinger will stop reporting latency to the monitor address and switch to reporting latency for the gateway address. Very obvious if the gateway address is local (~ 1ms) and the monitor address is across a remote link (~10ms). Once the switch occurs, apinger appears to remain stuck reporting the lower value.
It's notable that (hours) after the event, the pings are going to the correct (monitor) address even though the latency being reported corresponds to the local gateway address.
It's unclear what the trigger for this is. I haven't caught the event live yet. I have a tcpdump running in hopes of catching the event live.