A clock spike can occur due to packet collisions and network congestion. The NTP clock filter algorithm selects the most appropriate sample state and rejects noise spike. The clock offset and round-trip delay samples are computed from the four most recent timestamps based on the assumption that the network path is idle. If an offset sample exceeds the step threshold value, it is ignored and a timer is started. If a later sample is below the step threshold value, the counter is reset and the operation continues normally.
Because a spike is caused by network congestion over a path to a system peer, check network utilization and round-trip time over the network segments leading to the server using traceroute or ping, and also check system CPU status, which can cause the possible delay of NTP packet processing via the top command.
Info
Clock spike %val% seconds was detected while clock was synchronized with system peer.
Name | Type |
---|---|
val | String |