check_nt vs check_nrpe

[Flava]

Hi,

I have been reading around the NSClient++ agent used by NagiosXI to monitor Windows servers; the nsclient++ developers recommend using check_nrpe over check_nt, however I have noticed that monitoring wizard for Windows server use check_nt, What is the prefered plugin to use with NSClient++? And what is NagiosXI recommendation as best practise for configuring disk space monitoring? If check_nrpe is recommened plugin, can you please update monitoring wizard please?

Thank you

[scottwilkerson]

There isn't a "recommended" check type. It really depends on what you are trying to monitor on the machine.

check_nt offers easy ability to monitor windows performance counter

Whereas check_nrpe offers the ability to add custom scripts or [NRPE HANDLERS], which could be powerful for other types of checks.

[ahmad.zuhd]

What about the wizard for windows nrpe instead of check_nt ... is there a plan?

[scottwilkerson]

We have a basic NRPE wizard than can be used for any NRPE command. With windows NRPE, there aren't a ton of predefined checks, a wizard other that the standard NRPE wizard would need to know in advance what the name of the defined checks are.

[ahmad.zuhd]

All the predefined checks of the check_nt are also available for nrpe.... so can you plan to have a wizard for windows nrpe.

check the link
http://www.nsclient.org/nscp/wiki/CheckCommands

[slansing]

Those are the modules which control the NSClient++ commands and hand them off to NRPE to be sent to the Nagios server. I'm afraid I do not understand the following "you plan to have a wizard for windows nrpe." As we already have a NRPE wizard, and the Windows Desktop/Server monitoring wizard uses NRPE.

[notoncall]

As we already have a NRPE wizard, and the Windows Desktop/Server monitoring wizard uses NRPE.

I realize this is a dated conversation but it's the only thread I've found about this question.

I to am very pleased to have a wizard for Windows servers. This will make it much easier for my team to deploy monitors on their own.

However, when I use the Windows Server wizard, it uses the check_nt (nsclient) protocol on port 12489

In truth I want to use the NRPE protocol on port 5666.

The NSClient++ is already installed on the windows servers and the setup works fine in Nagios core.

I consulted the Wizard php code and indeed it uses nsclient protocol as seen in the example check command here:

check_xi_service_nsclient!!USEDDISKSPACE!-l D -w 80 -c 95

Here is the error I get: "connect to address 10.1.8.27 and port 12489: Connection refused"

Is there somewhere in the Windows server wizard that I can explicitly set that I want to use NRPE over port 5666? Or does the wizard only setup with check_nt (nslcient) over port 12489?

The NRPE Wizard is catered to unix hosts and doesn't have the nice features of the Windows Wizard for services, procs, drives, etc.

Many thanks.

[slansing]

You would need to go to Configure > CCM > Commands and duplicate the 'check_xi_service_nsclient' command, then change the port it uses there, and apply that command to your services. If you plan to use 5666 for every windows system you can just replace that port number in the original command I noted. Keep in mind that it will use this command for every windows system you set up through the Windows Server/Desktop wizards. You will also need to make sure that windows system is using 5666 for the nsclient protocol port and not 12489, which is default. You're going to want to make sure NRPE on those systems is not using 5666 anymore either otherwise you will run into socket lease issues if you are running both nrpe and check_nt checks to that system.

[notoncall]

Thanks for the speedy response! I think I understand you. The solution you proposed is to change the port number that the check_xi_service_nsclient command uses. That's not quite what I want to accomplish as that will continue to use nsclient (check_nt) protocol over the NRPE port 5666.

check_nt and check_nrpe have different command syntax too.

check_xi_service_nsclient: $USER1$/check_nt -H $HOSTADDRESS$ -s "$ARG1$" -p 12489 -v $ARG2$ $ARG3$ $ARG4$

check_nrpe: $USER1$/check_nrpe -H $HOSTADDRESS$ -t 30 -c $ARG1$ $ARG2$

What I really would like is for the Windows Server wizard to use the check_nrpe command and not the check_xi_service_nsclient command.

All of my existing checks use the NRPE protocol over port 5666 and using the check_nrpe command syntax like this:

check_nrpe -H 10.1.8.27 -c CheckDriveSize -a ShowAll MinWarnFree=10% MinCritFree=5% Drive=D:

You mentioned that the Windows server wizard uses NRPE. Maybe I'm simply missing the option for this method.

[lmiltchev]

You are correct. The "Windows Server" wizard uses "check_nt" by default. However, you can easily add more checks that use "check_nrpe" if you wish.
I am not sure when/if we are going to replace "check_nt" checks with "check_nrpe" checks in the wizard as were are gradually moving away from nsclient++. We decided to focus on our own agent - NCPA, which is a multi-platform agent. I suppose, most of the development efforts will be geared toward improving NCPA.