Page 1 of 2

Best Practices Questions

Posted: Thu Jun 25, 2015 1:44 pm
by daveinvb
I am building a Nagios XI environment from the ground up and I'm running into various issues and I'm trying to implement best practices in hopes to limit them.
I'm building it entirely in AWS, with 3 instances running XI, Fusion, and Network Analyzer and using AWS RDS for the database.
The basic layout of the environment is:
Image
This layout is replicated at nearly 50 locations with over 1,000 total host objects.
Each host object will have 1 Infrastructure Host Group and 1 Location Host Group.

So the first issue I'm running into is how to treat each host group differently. I want different check intervals and notifications for each type of infrastructure group. However, I'm having difficulty finding any information on how this is controlled and what inherits what from what. I can configure it for each object, I can configure it in a template, I can assign that template to a host or a host group or even another template. Case in point is that for 1 host, it says the template is "xiwizard_switch_host", which has the template "xiwizard_generic_host", which has the template "generic-switch", which is assigned to itself. The host group the host is assigned to has a separate template assigned to it. It's like chasing a problem down a rabbit hole that just gets deeper. What is the best way to accomplish this and what is the hierarchy?


The second issue I have is that the hosts I update say "Not Applied" after I apply changes. Looking through the forums, it says that this is generally a time issue. As far as I can tell the time settings are correct in both Nagios and on my server.

And lastly, I am having issues with slowness as the day progresses. After running for a couple days, it is unbearably slow.

Any suggestions would be greatly appreciated.

Re: Best Practices Questions

Posted: Thu Jun 25, 2015 2:23 pm
by tmcdonald
Let's tackle one thing at a time, since it gets really hard to deal with multiple issues in one thread. Please open separate threads for the slowness and Not Applied issues.
daveinvb wrote:I can assign that template to a host or a host group or even another template [...] The host group the host is assigned to has a separate template assigned to it.
I think this is your problem. This is a common misconception that leads many people to misconfigure. Templates are never applied to a hostgroup. Never ever ever. Ever. When you edit a host template (or a service template) and you click the "Manage Hostgroups" button, you are not applying that template to all hosts in that hostgroup, you are saying that any host using this template should belong to that hostgroup. Big difference.

Also, since you can attach a single service to a whole hostgroup (which means that service should be attached to all hosts in that hostgroup) you can very easily set up two sets of services, one set for each hostgroup, and give different intervals to each set.

With that in mind, do things make a bit more sense about how to configure this?

Re: Best Practices Questions

Posted: Thu Jun 25, 2015 2:29 pm
by eloyd
You can also hire someone like us to do it for you. :-)

Re: Best Practices Questions

Posted: Thu Jun 25, 2015 3:15 pm
by daveinvb
Trevor, thank you for your reply and I'll post the other questions in separate posts when I can get this one resolved.
So let me see if I understand. If I have a Host Template "server" and in the Host Template Management I click on Manage Hostgroups and I assign the groups "windows-server" and "unix-server" to that template, any host I then add the template to will be assigned to both host groups? If that's the case, that would definitely explain some things I've run into.

Right now, I'm just trying to set up a "host alive" basic ping and display up or down. We do have other services that we will need to monitor (cpu usage, sql services, memory, disk space, etc), but that's later. If I set this up as services connected to the host group, wouldn't that only display in the services dashlets? If I want it to be the hosts, will I need to create an individual template for reach group and manually assign it to each host in that group? Would that also be how I configure who to notify in each template?

Eric, unfortunately our Go Live is next week. But if you ever need someone to help with cloud work, let me know :D

Re: Best Practices Questions

Posted: Thu Jun 25, 2015 4:59 pm
by tmcdonald
daveinvb wrote:So let me see if I understand. If I have a Host Template "server" and in the Host Template Management I click on Manage Hostgroups and I assign the groups "windows-server" and "unix-server" to that template, any host I then add the template to will be assigned to both host groups? If that's the case, that would definitely explain some things I've run into.
Correct.
daveinvb wrote:If I set this up as services connected to the host group, wouldn't that only display in the services dashlets?
Can you clarify what you mean by this?
daveinvb wrote:If I want it to be the hosts, will I need to create an individual template for reach group and manually assign it to each host in that group? Would that also be how I configure who to notify in each template?
I think you've got the right idea, but I want to see your clarification above before I say for certain.

Re: Best Practices Questions

Posted: Fri Jun 26, 2015 8:44 am
by daveinvb
I haven't set up any dashlets yet, but the tactical overview is an example.

Image

I want all of the physical infrastructure to basically indicate if it is up or down. However, I also want to know services such as SQL running on some of those hosts.

So, to sum up....
For each Host Group, I should create a specific host template and manually assign it to each member in that group. This will allow me to monitor the group, ie "Wireless Access Points" as its own group, but also assign the members with their own check interval and allow me to modify them all at once.
I should also create a service template for the services I will monitor, such as "C Drive Space", and assign that to both host groups "windows servers" and "unix servers". That will allow me to monitor that service on those hosts but will not change their host groups.

Is that correct?

Re: Best Practices Questions

Posted: Fri Jun 26, 2015 8:50 am
by eloyd
Have you looked at some of the maps? Like the Minemap or BBmap? They may give you a better idea of what's up/down

Re: Best Practices Questions

Posted: Fri Jun 26, 2015 9:09 am
by tmcdonald
daveinvb wrote:I should also create a service template for the services I will monitor, such as "C Drive Space", and assign that to both host groups "windows servers" and "unix servers".
Assuming you mean the service itself when you say "assign that to both" then yes.

Re: Best Practices Questions

Posted: Fri Jun 26, 2015 9:25 am
by daveinvb
So I assign the service template to the service and then manually assign the service to each host? If I'm manually adding the service to the host, is there any reason to use a service template or is that just adding another layer of complexity when I can just make the changes to the service?

Is there any bulk add/remove for templates? The Bulk Modifications tool doesn't list templates.

Re: Best Practices Questions

Posted: Fri Jun 26, 2015 9:40 am
by tmcdonald
daveinvb wrote:So I assign the service template to the service
Yes
daveinvb wrote:and then manually assign the service to each host?
No, you would still assign the service to the host group.

daveinvb wrote:If I'm manually adding the service to the host, is there any reason to use a service template or is that just adding another layer of complexity when I can just make the changes to the service?
In this particular case, it is another layer of complexity. Templates are most useful when you have multiple hosts/services that use the template, so you can just make the changes in one place (the template) instead of many (each and every host/service that uses it).

daveinvb wrote:Is there any bulk add/remove for templates? The Bulk Modifications tool doesn't list templates.
This appears to have been added in a development revision, so it should be available in the next release.