Nagios Support Forum

Hi there folks

I'm currently evaluating NagiosXI and had hoped that it would fill all my needs. Have had it installed for a while but haven't had the time to conf it until now when I've had a few days away from work. I have to say, I'm a bit surprised over how complicated and tedious it is. I'm writing this post in the hope that I have missed something and that you can point me in the right direction.

The environment I want to monitor isn't that complicated. It consists of:

* Few OSX clients/servers
* Few switches, mellanox and HP
* 2 x Synology NAS
* HW Group poseidon 3266 with four sensor connected. 1 humidity and 3 temp.
* 2 x APC NetBotz 200
* 2 x APC ACRC301S, in-row coolers
* 16 x APC PDU AP8981
* 1 x APC Smart-UPS VT 40 kVA
* 1 x DDN NAS solution with 3 chassis
* 1 x Postgress database
* 1 x ftp
* 1 x http

Many of these I can monitor out of the box. Other's like the APC machines I have to resort to third party scripts for. And that's my first encounter with exchange.nagios.org. Was happy to see that there were loads of scripts for APC hardware but just as sad to see that most was abandoned skeletons that once worked or that just works for EOL models.

I have installed three scripts that works with the UPS (over snmp) to various extent. But there's really no way to in the GUI easily configure and manage graphs from what I can see? I have to set that up with templates that I have to script/code myself...? I would have expected a GUI for building graphs and relations.

When investigating graphing I tested nagiosgraph, since one apc script came with templates for it, but after a day of wrestling with it I disabled it again. When active it broke the regular graphs and either there's not a good installation guide or I'm stupid and incapable of reading and of basic understanding. Not sure if it's abandoned or not, but almost feels like it when looking at date's for releases? Same with PNP4Nagios.

For my ACRC301 in-row coolers there's no working script available that I can find. Looked at how to fix my own. But jesus christ that seems like a big endeavour. Sure, I can SNMP walk a device but to take that output and put it into the blender with MIB & OID's and then make something with perl that works with Nagios... Jeez. Tried the snmp walk wizard with the hope that it would give me an easy GUI to build something I could monitor. Couldn't find a way to force feed it just the APC MIB so what it found and the fields it populated didn't seem to make sense.

While sobbing over my failures here I instead looked at how to configure SNMP traps. Because that' a nice thing to have. But hey, the only trap here was the SNMP trap wizard. From what I can see the wizard does 0.4% of the setup and for the rest I once again have to study MIB's, OID's and endless of documentation. =/

Yes, I understand that this is a powerful front end to loads of misc tools and that this makes it very powerful. But at the same time, for a novice, it makes it into an unmanageable beast. And, yes... You might argue that "this *** is for grown up with huge unix beards, step aside puny n00b!". And you might very well have a good point there. But should it really have to be that way...?

There's a few things I fail to understand here. I'm not asking as an ***, I'm asking because I feel lost and I genuinely wonder.

* Why should I pay a premium for NagiosXI instead of using one of the various forks/clones or free solutions. Feels like I have to put in a ton of work no matter what solution I go for...?
* Have I missed something regarding my setup or is it really this tedious and complicated to get your devices monitored?


Cheers and thanks.


/Henrik

Moderator's Note: I had to edit your post a bit. Please use respectable language on our forum. Thank you!

Fortunately, I think you're right and that you just missed some things. I'll try to address as many points as I can:

NEO wrote:Other's like the APC machines I have to resort to third party scripts for. And that's my first encounter with exchange.nagios.org. Was happy to see that there were loads of scripts for APC hardware but just as sad to see that most was abandoned skeletons that once worked or that just works for EOL models

The Nagios Exchange is our site, but the intent of the site is to allow third-party developers and users to submit their own projects. As such, the majority of the code on the site is not our own. For this reason there might be some projects that have been discontinued, or that might not support newer devices. It is our preference to keep the projects listed so that they can be of some use, even if that means needing to modify them. And if the author does decide to return, the project is still there.

NEO wrote:But there's really no way to in the GUI easily configure and manage graphs from what I can see? I have to set that up with templates that I have to script/code myself...? I would have expected a GUI for building graphs and relations.

NEO wrote:When investigating graphing I tested nagiosgraph, since one apc script came with templates for it, but after a day of wrestling with it I disabled it again. When active it broke the regular graphs and either there's not a good installation guide or I'm stupid and incapable of reading and of basic understanding. Not sure if it's abandoned or not, but almost feels like it when looking at date's for releases? Same with PNP4Nagios.

Graphs are built automatically now in XI and have been for a while - no need to configure templates or anything. Judging by your comments about Nagiosgraph and PNP4Nagios, you likely were reading Core documentation as opposed to XI documentation. Installing additional graphing software on an existing XI system can absolutely cause some problems. In addition, neither Nagiosgraph nor PNP4Nagios were written by Nagios, so we don't control their codebases.

NEO wrote:For my ACRC301 in-row coolers there's no working script available that I can find. Looked at how to fix my own. But jesus christ that seems like a big endeavour. Sure, I can SNMP walk a device but to take that output and put it into the blender with MIB & OID's and then make something with perl that works with Nagios... Jeez. Tried the snmp walk wizard with the hope that it would give me an easy GUI to build something I could monitor. Couldn't find a way to force feed it just the APC MIB so what it found and the fields it populated didn't seem to make sense.

While we do try to provide out-of-the-box plugins for most of what people monitor, there's no way we can include a plugin for each specific make and model of device. The SNMP Walk Wizard is a great choice for monitoring snmp-capable devices, and there is no need to write a plugin specifically for that device in most cases. In regards to the MIB, there is a page located under Admin -> Manage MIBs that can import the MIB files. Off the top of my head I do not know which of our wizards will respect those values - I honestly would need to test or check with the developers on this.

NEO wrote:While sobbing over my failures here I instead looked at how to configure SNMP traps. Because that' a nice thing to have. But hey, the only trap here was the SNMP trap wizard. From what I can see the wizard does 0.4% of the setup and for the rest I once again have to study MIB's, OID's and endless of documentation.

The SNMP trap documentation is pretty simple - https://assets.nagios.com/downloads/nag ... ios_XI.pdf

But yes, some manual configuration needs to be done so Nagios knows what a particular trap variable means in the context of your environment.

NEO wrote:And, yes... You might argue that "this *** is for grown up with huge unix beards, step aside puny n00b!". And you might very well have a good point there. But should it really have to be that way...?

We designed XI specifically so people don't need to grow huge Unix beards. 95% of what needs to be done on an XI system can be done from the web interface. There are some specific things that must still be done manually, most of them having to do with permissions or security, but a vast majority of the day-to-day is done right from the web console.



I hope that helps clear up some of the misconceptions! We try to make XI as straight-forward as possible, but if you need some help getting off the ground we offer a free, one-hour session called a Quickstart: https://www.nagios.com/services/quickstart/

You fill out a short questionnaire, then pick a time on the calendar. One of our technicians will send you connection information shortly before the session is scheduled to begin. We use WebEx or GoToAssist, and it is strongly advised to have a phone so we can chat via voice and not just text.

Hello!

Thanks for your reply! =)

tmcdonald wrote:Fortunately, I think you're right and that you just missed some things. I'll try to address as many points as I can:

NEO wrote:Other's like the APC machines I have to resort to third party scripts for. And that's my first encounter with exchange.nagios.org. Was happy to see that there were loads of scripts for APC hardware but just as sad to see that most was abandoned skeletons that once worked or that just works for EOL models

The Nagios Exchange is our site, but the intent of the site is to allow third-party developers and users to submit their own projects. As such, the majority of the code on the site is not our own. For this reason there might be some projects that have been discontinued, or that might not support newer devices. It is our preference to keep the projects listed so that they can be of some use, even if that means needing to modify them. And if the author does decide to return, the project is still there.

Gotcha. I think what's missing there is a clearer overview of when something have been changed and with what it actually works. But guess that historically have been up to the author to add and that it therefore is a bit fuzzy on some items.

tmcdonald wrote:

NEO wrote:But there's really no way to in the GUI easily configure and manage graphs from what I can see? I have to set that up with templates that I have to script/code myself...? I would have expected a GUI for building graphs and relations.

NEO wrote:When investigating graphing I tested nagiosgraph, since one apc script came with templates for it, but after a day of wrestling with it I disabled it again. When active it broke the regular graphs and either there's not a good installation guide or I'm stupid and incapable of reading and of basic understanding. Not sure if it's abandoned or not, but almost feels like it when looking at date's for releases? Same with PNP4Nagios.

Graphs are built automatically now in XI and have been for a while - no need to configure templates or anything. Judging by your comments about Nagiosgraph and PNP4Nagios, you likely were reading Core documentation as opposed to XI documentation. Installing additional graphing software on an existing XI system can absolutely cause some problems. In addition, neither Nagiosgraph nor PNP4Nagios were written by Nagios, so we don't control their codebases.

Thanks for clearing this up. But if you look at the image here, what I cannot understand is how I, in the webgui, create a graph for the output of this apc-script. Can't find any function like that.

tmcdonald wrote:

NEO wrote:For my ACRC301 in-row coolers there's no working script available that I can find. Looked at how to fix my own. But jesus christ that seems like a big endeavour. Sure, I can SNMP walk a device but to take that output and put it into the blender with MIB & OID's and then make something with perl that works with Nagios... Jeez. Tried the snmp walk wizard with the hope that it would give me an easy GUI to build something I could monitor. Couldn't find a way to force feed it just the APC MIB so what it found and the fields it populated didn't seem to make sense.

While we do try to provide out-of-the-box plugins for most of what people monitor, there's no way we can include a plugin for each specific make and model of device. The SNMP Walk Wizard is a great choice for monitoring snmp-capable devices, and there is no need to write a plugin specifically for that device in most cases. In regards to the MIB, there is a page located under Admin -> Manage MIBs that can import the MIB files. Off the top of my head I do not know which of our wizards will respect those values - I honestly would need to test or check with the developers on this.

I got the tip to test Observium and the way they handle this is pure magic. It "walks" the device and adds, from what I can see, everything. I don't have to do anything. I've attached a screenshot from what Observium did after polling my devices. Took me less than an hour including install of host OS (VM). After base install there was no fiddling around in the terminal. Not saying that Observium is better than Nagios or anything, still evaluating both. I'm sure Observium will have it shortcomings but when it comes to device discovery and support this is pretty much what I'd expect from an enterprise solution I have to pay for.

tmcdonald wrote:

NEO wrote:While sobbing over my failures here I instead looked at how to configure SNMP traps. Because that' a nice thing to have. But hey, the only trap here was the SNMP trap wizard. From what I can see the wizard does 0.4% of the setup and for the rest I once again have to study MIB's, OID's and endless of documentation.

The SNMP trap documentation is pretty simple - https://assets.nagios.com/downloads/nag ... ios_XI.pdf

But yes, some manual configuration needs to be done so Nagios knows what a particular trap variable means in the context of your environment.

Yeah, I've read that trap document. And to get it to work I have to dig out OID's manually and add them to snmptt.cfg. Maybe this is something others can wake up and do in the middle of the night. But to even find the correct OID that corresponds to what I want to do feels like a monumental task on its own. Not saying that this is a nagios issue per say since I'm well aware that you're using standards tools in the background. What I miss is out of the box configuration wizards that takes care of all terminal code hacking and all tedious tasks for me. A wizard that really is a wizard and that does magic.

tmcdonald wrote:

NEO wrote:And, yes... You might argue that "this *** is for grown up with huge unix beards, step aside puny n00b!". And you might very well have a good point there. But should it really have to be that way...?

We designed XI specifically so people don't need to grow huge Unix beards. 95% of what needs to be done on an XI system can be done from the web interface. There are some specific things that must still be done manually, most of them having to do with permissions or security, but a vast majority of the day-to-day is done right from the web console.



I hope that helps clear up some of the misconceptions! We try to make XI as straight-forward as possible, but if you need some help getting off the ground we offer a free, one-hour session called a Quickstart: https://www.nagios.com/services/quickstart/

You fill out a short questionnaire, then pick a time on the calendar. One of our technicians will send you connection information shortly before the session is scheduled to begin. We use WebEx or GoToAssist, and it is strongly advised to have a phone so we can chat via voice and not just text.

Cool. I had forgotten about that quickstart! Will make use of that once I'm back at work. Think it will be valuable. Still evaluating it and will test a few more solutions before deciding on what route to take. When it comes to device discovery and autoconfig it seems like Observium is in a class of its own though. It'll probably have other shortcomings though. =/

Cheers and thanks

NEO wrote:Thanks for clearing this up. But if you look at the image here, what I cannot understand is how I, in the webgui, create a graph for the output of this apc-script. :roll: Can't find any function like that.

The graphs should be present in that second tab, the one that, to me, looks like a little mountain. Just to the right of the Overview tab.

NEO wrote:I got the tip to test Observium and the way they handle this is pure magic. It "walks" the device and adds, from what I can see, everything. I don't have to do anything. I've attached a screenshot from what Observium did after polling my devices. Took me less than an hour including install of host OS (VM). After base install there was no fiddling around in the terminal. Not saying that Observium is better than Nagios or anything, still evaluating both. I'm sure Observium will have it shortcomings but when it comes to device discovery and support this is pretty much what I'd expect from an enterprise solution I have to pay for.

It looks like they put a lot of their focus on autodiscovery. Unfortunately according to their documentation it appears that SNMP is a requirement: http://www.observium.org/docs/add_device/

I'll admit that, without having tried it myself, their autodiscovery looks like it's pretty nice. But at the cost of limiting yourself to only SNMP checks I can't say I would ever consider that to be a net positive.

NEO wrote:Yeah, I've read that trap document. And to get it to work I have to dig out OID's manually and add them to snmptt.cfg. Maybe this is something others can wake up and do in the middle of the night. But to even find the correct OID that corresponds to what I want to do feels like a monumental task on its own. Not saying that this is a nagios issue per say since I'm well aware that you're using standards tools in the background. What I miss is out of the box configuration wizards that takes care of all terminal code hacking and all tedious tasks for me. A wizard that really is a wizard and that does magic.

Without having set up Observium before, I can't really speak too much to how they do things. I tend to believe there was more to it than just point-and-click since the MIBs need to be uploaded somehow, but I digress. We do a lot of things well, and we have a ton of wizards that do seem to work like magic, but some areas will need some time to be taken before the magic can happen.



There are plenty of ways to do what you need using Nagios, and some will take more effort than others. But overall, even considering the setup that needs to be done, I'd say you are getting much more benefit from Nagios with the graphing, reporting, alerting, in-browser configuration, backups, etc. etc.