Changing from ncpa to MOD-Gearman, good idea?

[Toby]

Hi again!

Not sure this is the right forum index, but here goes
I have 1 Nagiosxi server.
I also have about a dozen windows Machines, running NCPA agent and reporting in to Nagiosxi.

There is a will to increase the monitoring from "awkward" sites, measuring the quality of our delivery.
That means probably about 20 nodes more.
Good thing is that we have a whole lot of Lenovo thin clients laying around, so hardware is almost free.
I'm planning to install Centos and then MOD-Gearman to monitor our stuff.
My question is:
You think that is the right way to go?

My 5 cents.
pros:
cheap Machines, will have enough juice.
Centrally controlled by nagiosxi

Cons:
Have to downgrade xi? How will updates work in the future then?
Is MOD-Gearman still an Active/supported Project?

Is there a better alternative that i missed out on?

[lmiltchev]

You could use ModGearman if you really need to... You won't need to downgrade Nagios XI, but Nagios Core. For more information on the topic, review the KB article below:

https://support.nagios.com/kb/article/n ... e-823.html

There shouldn't be a problem with upgrading XI in the future. Our upgrade script checks to see if you are actually using ModGearman. If you are, then Nagios Core doesn't get upgraded. Also, our developers will be working on ModGearman to make it compatible with Nagios Core 4.4.x, even though this may take some time. I don't have an ETA on that.

Having said that, I would recommend that you don't use ModGearman unless you have to. Even though ModGearman is a great software, it is just one more "moving part" that complicates things.

What kind of resources do you have on the Nagios XI server, and how many hosts/services are you monitoring? There are many ways to tweak XI, and reduce the load on the server (besides using Gearman). For example, you could use mostly passive checks with NCPA, and have the results sent to your Nagios XI server.

Hope this helps.

[Toby]

Hi!

Thx for a fast reply.

"then Nagios Core doesn't get upgraded. Also, our developers will be working on ModGearman to make it compatible with Nagios Core 4.4.x"
Great news!


"What kind of resources do you have on the Nagios XI server, and how many hosts/services are you monitoring? "
I do not use ncpa for offloading the Nagiosxi server.
My Nagiosxi server is ridiculously overpowered.
Physical server hp380p gen 8 with 2*8 cores(+HT), fully smacked with ssd's in raid 1 and 5, and 256 gb ram.
(just had it laying around ) About 150 Hosts, 400 services.

We are mostly using Nagiosxi to check the quality of our delivery from the user perspective.
My Nagios does not check all nodes in a web-cluster, we check log in and page visits as the user do. my check is not interested if 1 node is down in the cluster. Little like Webinject in nagiosxi, but i use powershell atm.
Clusterservices and stuff are checked with other stuff.

Our small Machines with ncpa are placed in small branch offices, that is placed all over the county(Health centers).
They are measuring the quality from the users location, not a cold server room, very close to the server delivering the service.
And as a secondary benefit, they are very good for all kinds of troubleshooting

What i need is a more stable plattform than small windows -Machines and ncpa. And a more centralized management.
From what i read, i will get that from Gearman.

I hope you understand my explanation of our setup, or i will have to doodle in paint

[scottwilkerson]

I hope you understand my explanation of our setup, or i will have to doodle in paint

I understand what you are saying, but I'm not so sure you are making things less complicated by introducing mod_gearman as a matter of fact, I believe you would be increasing the complexity of your setup 20 fold by adding mod_gearman workers vs just having NCPA on the remote machines.

With a mod_gearman setup like you propose, you would have to define countless different groups and only have each worker process the results of each special group. Then you would need to configure each check to be part of a specific group when configuring /modifying it...

If you setup a similar setup with NCPA and passive checks on each, you simple need to setup one of the configs just the way you like it, and then copy the same configuration to each of your hosts and they will all start reporting back to your Nagios server.