Nagios XI for Switches and Routers

Support forum for Nagios Core, Nagios Plugins, NCPA, NRPE, NSCA, NDOUtils and more. Engage with the community of users including those using the open source solutions.
Locked
DFaught
Posts: 62
Joined: Tue Sep 26, 2017 12:50 pm

Nagios XI for Switches and Routers

Post by DFaught »

Using Nagios XI for network devices, by which I mean routers and switches, is kludgy. Because I see XI wizards for both straight SNMP service monitors and for MRTG-based service monitors with the MRTG-based services being the preferred method, I assume that someone has done a comparison of using both methods and decided that the MRTG monitoring has a more efficient SNMP polling engine or is better in some other important way. If that kind of comparison has not been done, then I have to wonder why MRTG was ever introduced into the Nagios mix. If MRTG is really a better, more efficient method, then there are several things that I would like to see improved in the integration between Nagios and MRTG in order for Nagios to truly be an enterprise-class network monitoring system. In my case, enterprise-class means a wide variety and hundreds of network devices with thousands of ports.

1. One of the most disappointing things about trying to use Nagios XI for network devices is that the nice Nagios document "Automated Host Management in Nagios XI" does not work for network devices and their services because it does not extend to the MRTG part of the product. If I pursued doing this with straight SNMP then it would be a totally different configuration than what the Network Switch/Router Configuration Wizard produces, and following the logic above, possibly less efficient than using MRTG.

2. The combination of using "Monitor Using: Port's Description" in Step 1 of the Network Switch/Router Configuration Wizard and using the REST API "GET objects/rrdexport" query is very awkward. The long strings of the port descriptions are extremely desirable in many or most places in the XI GUI, but are clumsy in the REST API. Maybe a combination of using the port numbers internally and in the REST API, and either the port names or descriptions in the GUI would work well.

3. Cleaning up MRTG configurations when devices or services are deleted in the Nagios XI database is at least a recognized issue, although needing to do this manually is asking for trouble.

4. Another very disappointing thing about the Nagios XI/MRTG mindmeld is that it is virtually impossible to add other port statistics to the monitoring configurations. In particular it is very desirable to have another column or two in Step 2 of the Network Switch/Router Configuration Wizard to be able to optionally monitor input and output errors on switch and router ports. I do really like the column header Check/Uncheck options on that page.

5. I saw another post on this forum about adding a Rediscover button so that changes to port descriptions or speeds could be easily updated without going through the entire wizard over again. I like this idea too, although even better would be if this was just done automatically on some periodic basis.

Thank you for your kind attention.
bheden
Product Development Manager
Posts: 179
Joined: Thu Feb 13, 2014 9:50 am
Location: Nagios Enterprises

Re: Nagios XI for Switches and Routers

Post by bheden »

I assume that someone has done a comparison of using both methods and decided that the MRTG monitoring has a more efficient SNMP polling engine or is better in some other important way.
Your assumption is correct in this case.
1. One of the most disappointing things about trying to use Nagios XI for network devices is that the nice Nagios document "Automated Host Management in Nagios XI" does not work for network devices and their services because it does not extend to the MRTG part of the product. If I pursued doing this with straight SNMP then it would be a totally different configuration than what the Network Switch/Router Configuration Wizard produces, and following the logic above, possibly less efficient than using MRTG.
My intention is not to excuse this point, but I'll go on the record in saying that we're working on an overhaul of the bandwidth style monitoring for a future major release of XI. This won't be in there until XI 6 likely, because of many of the things you've addressed here.
2. The combination of using "Monitor Using: Port's Description" in Step 1 of the Network Switch/Router Configuration Wizard and using the REST API "GET objects/rrdexport" query is very awkward. The long strings of the port descriptions are extremely desirable in many or most places in the XI GUI, but are clumsy in the REST API. Maybe a combination of using the port numbers internally and in the REST API, and either the port names or descriptions in the GUI would work well.
I like this idea. We've actually toyed with something similar for maintaining RRD/service ID correlation from in XI, but ultimately the features we developed for this never made it into the product. I think revisiting this could have desirable outcomes beneficial for all of us.
3. Cleaning up MRTG configurations when devices or services are deleted in the Nagios XI database is at least a recognized issue, although needing to do this manually is asking for trouble.
This is the "something similar" I mentioned in the previous point. This functionality is I'd wager half complete at this time. Not an excuse, just trying to be transparent with what's happening here. As stated already, revisiting this at least will be good for the short term before we do our bandwidth overhaul.
4. Another very disappointing thing about the Nagios XI/MRTG mindmeld is that it is virtually impossible to add other port statistics to the monitoring configurations. In particular it is very desirable to have another column or two in Step 2 of the Network Switch/Router Configuration Wizard to be able to optionally monitor input and output errors on switch and router ports. I do really like the column header Check/Uncheck options on that page.
In the past I've just kept my specific routers/switches in device-based hostgroups, and attached services to the hostgroups to solve some of these problem. Specific SNMP checks, pulling community/version data from free variables assigned to the host. And then scripting some automation to add any additional port checking. Obviously this isn't optimal by not being done for you out of the box.
5. I saw another post on this forum about adding a Rediscover button so that changes to port descriptions or speeds could be easily updated without going through the entire wizard over again. I like this idea too, although even better would be if this was just done automatically on some periodic basis.
I also like this idea.

I've added the following feature requests:

#13106: MRTG configuration files syncronized with CCM revisit
#13107: Keep track of both port number and description in MRTG based services
#13108: Add ability to monitor additional port metrics in the switch wizard
#13109: Rediscover option or continuously poll network devices for changes

These were all created with the target version of XI 6.0.0, so it may still be quite a while before you see them in action.

Thank you very much for your insight.
As of May 25th, 2018, all communications with Nagios Enterprises and its employees are covered under our new Privacy Policy.

Nagios Enterprises
Senior Developer
Locked