https://rancher.com/ logo
Title
l

late-alarm-70776

01/24/2023, 4:49 PM
Kubelet error log "Failed updating cpu usage nano core" Cluster information: Kubernetes version: v1.21.14+k3s1 In journalctl logs I am seeing lot of error occurrences with reason
NodeNotReady
following
Jan 24 05:40:16 xxxxx k3s[1162]: E0124 05:40:16.729337    1162 cri_stats_provider.go:683] "Failed updating cpu usage nano core" err="zero or negative interval (1674538816724786490 - 1674538895342330208)"
logs. Similar issue is reported by another user at https://discuss.kubernetes.io/t/kubelet-error-log-failed-updating-cpu-usage-nano-core/22134 Please let me know what issue can cause this problem and how to address the same.
Answering it myself. After investigating bit more this looks like intermittent NTP sync issue due to which Kubernetes errorred out while computing CPU usage based on the current and the cached time and K8s at that time transitions node into
NodeNotReady
state and reschedule all the pods.