[ntp:questions] nagios

Rob nomail at example.com
Tue Jun 10 07:58:50 UTC 2014


Rob <nomail at example.com> wrote:
> Harlan Stenn <stenn at ntp.org> wrote:
>> Rob writes:
>>> I am trying to monitor NTP servers with nagios.
>>> I use the standard plugin check_ntp_peer.
>>> 
>>> It appears that it is only checking the offset and jitter of the
>>> currently selected peer.  When I run check_ntp_peer with some threshold
>>> it returns the same values as present on the ntpq output line with *
>>> at the beginning.
>>> 
>>> I think this is wrong.  It should do the equivalent of "ntpq -c rv"
>>> and pick the offset and sys_jitter values from there.
>>> 
>>> Is anyone aware of a nagios plugin that does that?
>>
>> Have you seen http://support.ntp.org/Support/MonitoringAndControllingNTP
>>
>> and the check_ntpq nagios module it mentions that I haven't touched in a
>> Long Time?
>>
>> H
>
> Ok I see that this one at least uses "ntpq -c readvar".
>
> I would prefer an implementation that uses the readvar packet itself,
> as I think that it can return more precision than ntpq does.
> (its float value formats are limited to 3 digits after the decimal)

Well, I now see that the readvar packet returns the data in the same
format (as ASCII string), not as a fixed point value.  And I realize
that ntpd operates in units of a microsecond anyway, not a nanosecond
as the kernel itself.

Most of the systems I want to monitor are synced to PPS, so the output
of the readvar command is usually something like:

offset=-0.001, sys_jitter=0.001

It would have been nice if this was detailed a bit more, but apparently
this is not going to happen.

I downloaded your perl plugin and made some modifications to use it
with the current ntpq output.  Let's see how it goes.



More information about the questions mailing list