Thanks for the response Brian. I have already enabled the NTP collector in all all my servers, but still cannot see the *node_ntp_drift_seconds* metrics giving the output. Apart from that, I have couple of questions here. Firstly, why are we checking the target clock with Prometheus' server? What if it itself get unsyncronized? The whole idea of alerting goes out of the water in that case. Also, what does the node_ntp_sanity checks? How much is the variation in the clock that it takes into consideration to make the sanity 0(I know other factors also can make the sanity 0, but what is the criteria to call the clock *unsynchronized*. Same question for node_ntp_leap, if leaps turn 3, that means it is unsynchronized. Again what is the difference in clock timings that is takes to call the clock *unsynchronized*?
Secondly, according to you which one is better for keeping a track of clock Sync? Timex or NTP? On Tuesday, May 19, 2020 at 2:32:06 PM UTC+5:30, Brian Candler wrote: > > The ntp collector is disabled by default > <https://github.com/prometheus/node_exporter#disabled-by-default>: you > can turn it on with a command-line flag. However, the timex collector is > enabled by default (e.g. node_timex_sync_status, > node_timex_estimated_error_seconds) > > For a rough idea of how the target clock compares to the prometheus > server's clock, you can also just do: > > node_time_seconds - timestamp(node_time_seconds) > > -- You received this message because you are subscribed to the Google Groups "Prometheus Users" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/prometheus-users/ac52c604-53cd-4a20-a7fd-7b46e2a8bf65%40googlegroups.com.
