Page 1 of 1

Check_nrpe output

Posted: Thu May 08, 2014 5:34 am
by Narie
Hi,

I've got a Windows server which is monitored by a NRPEproxy. The NRPE line in nrpe.conf is:
command[check_drive-servername]=/usr/lib/nagios/plugins/check_nrpe -H 10.10.10.10 -c CheckDriveSize -a Drive="C" MaxWarnUsed=89% MaxCritUsed=90%
In NagiosXi I have created a new service with a check_nrpe command and as argument "check_drive-servername". When I do a test (test Check Command) and I fill the NRPE proxy i receive an complete line:
OUTPUT: OK: All drives within bounds.|'C: %'=88%;89;90 'C:'=8.82GB;8.89;8.99;0;9.99 (the same as when I try on the command line of the proxy host. However, the output in NagiosXi Service Status is taken down in OUTPUT: "OK: All drives within bounds.". I am only interested in the second part after the pipe.

Re: Check_nrpe output

Posted: Thu May 08, 2014 9:27 am
by slansing
So first off, are you using win_nrpe? That has been depricated for years, the current recommended windows agents are NCPA or NSClient++. Secondly, the data after the pipe is performance data that is processed and routed to RRD's which in turn are used to populate your graphs. You can see this data by going to the advanced tab of your service:

https://nagios-plugins.org/doc/guidelines.html#AEN200

Re: Check_nrpe output

Posted: Fri May 09, 2014 3:39 am
by Narie
We are talking about the NSClient++.
The difference between other Windows hosts is the NRPE proxy. How can I simple get the same status information ( D:\ - total: 2.02 Gb - used: 0.54 Gb (27%) - free 1.48 Gb (73%)) instead of "Ok: within bounds". For me this seems to bee an check_nrpe question?

Re: Check_nrpe output

Posted: Fri May 09, 2014 10:59 am
by sreinhardt
The behavior you are discussing is fully expected, the service details display message and performance data are considered separate, and happen to be separated via a pipe. Since the performance data is only used for graphing, unless you use $LONGSERVICEOUTPUT$ on items like notifications, you will not get anything past the |. As for within the XI interface, There really isn't anything we can do unless you wrote a wrapper script that calls check_nt and reformats the output into something more usable for you, such as containing that disk perf data prior to the pipe also. Unless I missed something in your post, I would expect this to be the same for all of this type of check, no?