[ovs-discuss] Prometheus Exporter for OVN

Vasiliy Tolstov v.tolstov at selfip.ru
Fri Oct 5 20:34:05 UTC 2018


>
> Here is the thinking on this one. The presence of pid file and the validation that the process listed in the pid file are process checks. If the process is 0, then it is down and one can monitor for that. Also, if the process ID of a subsystem (say OVN NB) changed, then it means that service was restarted. I guess the alert would be if a process is 0 or if the process ID changed, then alert.
>

Does it possible to monitor ovn service uptime? If uptime bigger then
before - service is restarted.

> Per the log files. It happened on a number of occasions that a subsystem crashed, but its process was still running. One way of detecting something like that is watch for the log file being stale, i.e. log size does not increment. There are obviously other ways to monitor that. Also, any spike in log activity, i.e. sudden large increase in log file size, means something is happening and it worth looking into it.

I think that it is not right way. Sometimes i want to write logs to
syslog or journald...

>
> > also system run dir. May be to monitor disk
> space, but i'm confused.
>
> That's really informational. Kinda know where system dir is.

For what this mertic can be used? If i have debian and fedora systems
and this dir may changed over distributions....


-- 
Vasiliy Tolstov,
e-mail: v.tolstov at selfip.ru


More information about the discuss mailing list