gearman / mod_gearman -> NCPA

[invade]

Hi.

We use Nagios to monitor a few hundred systems that are deployed to customer sites.

None of the sites have inbound access allowed and so we use mod_gearman (send_gearman) to run passive checks.

I have a number of bespoke scripts that check various things and then call send_gearman supplying the following:
Service Checks: <host_name>[tab]<svc_description>[tab]<return_code>[tab]<plugin_output>[newline]
Host Checks: <host_name>[tab]<return_code>[tab]<plugin_output>[newline]

The results go to a gearman server which Nagios Core then polls.

The gearman / mod_gearman software is deployed using packages created by Consol Labs (https://labs.consol.de/repo/stable/)

We are in the process of migrating to CentOS 8. I've been in touch with ConSol Consulting & Solutions Software GmbH to ask if they plan to release any packages for RHEL8 / CentOS 8 and have been advised "Currently we have no such plans. None of our customers is requesting this change, currently."

As a result I'm looking at alternatives methods for sending passive checks.

Can NCPA be used as a direct replacement for our gearman / mod_gearman based solution? eg. can I run custom passive checks using NCPA which are effectively the results of my bespoke scripts?

I am aware that neither gearman or mod_gearman are Nagios products. I'm hoping to replace these products with native Nagios functionality.

Any advice is much appreciated. Many thanks in advance.

[mbellerue]

NCPA can certainly be used to run custom passive checks. I'm having a little trouble wrapping my head around the setup you've described, though.

First, let's see if I'm understanding your setup correctly. Let's say you have a customer with 25 client machines and 1 machine checking the health of those 25 machines. The 1 machine reports back to the main Nagios XI instance with the health of the 25 client machines. Then multiply that by however many customers you have.

Does that sound like what you're doing? If that's the case, it might just make sense to have a Nagios Core instance on the customer network sending data back to XI via NRDP.

[cdienger]

NCPA is capable of sending passive checks via NRDP:

https://assets.nagios.com/downloads/ncp ... Checks.pdf

You'll then also need to install NRDP on the Core system:

https://support.nagios.com/kb/article/n ... e-602.html

[invade]

NCPA can certainly be used to run custom passive checks. I'm having a little trouble wrapping my head around the setup you've described, though.

First, let's see if I'm understanding your setup correctly. Let's say you have a customer with 25 client machines and 1 machine checking the health of those 25 machines. The 1 machine reports back to the main Nagios XI instance with the health of the 25 client machines. Then multiply that by however many customers you have.

Does that sound like what you're doing? If that's the case, it might just make sense to have a Nagios Core instance on the customer network sending data back to XI via NRDP.

Hi. Not quite. Each separate client system runs a bunch of bespoke scripts (the passive checks), each of which then send the results to a gearman server using the "send_gearman" mod_gearman utility. All the results from all the separate scripts, from all the separate client systems, are queued on the gearmand server. Nagios then polls the gearmand server using the mod_geaman NEB module and actions the results accordingly.

If possible, I would like to take gearman / mod_gearman out of the equation and have passive checks running on the client systems that send their results directly to Nagios.

I had a quick look at NCPA and it looked like it could do this, my only concern was that in the examples all the checks were fairly standard re-defined checks, eg. filesystem x occupancy is y, etc.

All the checks we run are fairly bespoke so, as long as I can pass the results of my scripts to NCPA in a format that can be used as a check results, I believe it should be OK.

[invade]

I've just discovered the Nagios repository:
https://repo.nagios.com/
(I didn't realise this exists, our Nagios installation is currently built from packages in the EPEL repository).

I notice that the el7 repository (https://repo.nagios.com/nagios/7/) contains both gearmand and mod_gearman. It also contains NCPA but not NRDP (that I can see).

The el8 repository (https://repo.nagios.com/nagios/8/) only contains NCPA.

Packages are our preferred method of software deployment so, I'm not sure who the packagers are for this repository but I now have two extra questions:

1. Will packages for gearmand and mod_gearman be created for el8?

2. Is there a reason (maybe licencing) why there are no packages for NRDP in any repositories, for any Linux distributions, that I can find?

Thanks again.

[cdienger]

I'll have to check with our devs regarding 1, but to answer your second question, NRDP is a protocol that NCPA(and other agents) are capable of using.

[invade]

I'll have to check with our devs regarding 1, but to answer your second question, NRDP is a protocol that NCPA(and other agents) are capable of using.

Thanks for the responses.

Apologies if I've misunderstood but, does your previous post not suggest that NRDP is a separately install-able component on the Nagios Core server? If this is the case, then it would be preferable to do this via a package to supplement the Nagios Core installation which has been deployed using packages. It just seemed odd to me that there was no package for NRDP in the Nagios Repos, the EPEL repo, or any other 3rd party repos that I could find where Nagios components were being packaged.

[cdienger]

Ah! Sorry, I thought you were referring the the clients. You're correct that we don't have an official repo for the server portion. I'm not aware of any plans to.

We're working on support for XI on 8 at the moment. No time frame available, but we hope to have the modgearman components available around the same time.

[invade]

Ah! Sorry, I thought you were referring the the clients. You're correct that we don't have an official repo for the server portion. I'm not aware of any plans to.

We're working on support for XI on 8 at the moment. No time frame available, but we hope to have the modgearman components available around the same time.

Thanks for the update. If you will be packaging up gearman / mod_gearman for el8 then I might hang fire rather than go through all the effort to migrate to NCPA at this time.

[cdienger]

Sounds good. Do you have any further related questions or are we okay to lock the thread at this point?