The first tool I evaluated was Spiceworks. This is a windows application that you can have installed and monitoring your network within 30 minutes or less. It has a beautiful GUI and you can set it to scan your network, discover your devices, classify them and start monitoring them very easily. It can monitor disks, network interfaces, printer toner, and even antivirus status on client machines. It has good reporting options and includes a help desk solution. You can use it free of charge. I tested it with about 50 devices and it worked without issues. It uses WMI, SSH and SMTP to monitor devices, so it doesn’t need an agent to be installed on monitored machines.

This tool is not open source. You can add some functionality through some kind of add-ons, but you can’t modify the core application. Additionally, they keep the tool free by including some ads in the monitoring console. These were the two main things I didn’t like about Spiceworks, and this is why I decided to evaluate some open source alternatives.

I evaluated 3 open source options: Nagios, Zabbix and Zenoss. I evaluated the free versions of these products (not Nagios XI nor Zenoss enterprise). At the end, we opted for Zenoss (core).

The first thing I should say, is that I think that any thing that can be done with any of these 3 tools, can be done with the other two as well, but probably in a different way and with more or less work. Additional functionality can be added to all 3 by means of plug-ins, add-ons or zenpacks, and as they are open source, you can modify the core application as well if you need to.

I post here the advantages and disadvantages I saw on each one, based on our infrastructure, requirements and other factors that could be particular for our organization. However, this might serve as a starting point for someone wanting to evaluate and choose a free and open source monitoring solution.

NAGIOS:

Advantages:
-It is the less resource-hungry one.
-Its web interface is very fast and simple.
-It has a large user base, you can easily find help.
-Acceptable documentation.
-Many plug-ins and ad-ons available.
-Very flexible
-Network mapping capabilities.
-Many advanced features (which are not visible through the web interface, but are configurable through the many configuration files it uses).

Disadvantages:
-Configuration changes must be done by directly modifying the configuration files. The default web interface does not include configuration options.
-It seems that there is no native support for authentication when sending e-mail alerts. One can certainly make it work by installing and configuring some additional tools, but it requires some work.
-I didn’t like the web interface too much.
-The default web interface does not include graphics that one can use to make a trending analysis, it shows only real time data. One can generate graphics with the help of some additional tool that needs to be installed and configured.
-The default reports included are very few.
-It does not offer information useful to inventory the monitored devices (I know, it is a monitoring tool, nothing more)
-Everything is stored on text files (not database).
-An agent needs to be installed on (most of) the monitored devices (having an agent could be an advantage in some circumstances).
-Performance graphs, configuration wizards, custom dashboards and other features are available only in the commercial version (Nagios XI).

ZABBIX:

Advantages:
-It offers graphing capabilities out of the box. Graphs can be generated for almost anything.
-It shows many details (hardware and software) about monitored devices.
-It has device discovering capabilities, and with the agent installed on the devices, the auto discovery process can classify them and assign them the proper monitoring template, automatically.
-Everything is stored in a database, so you can easily query the information to create custom reports or with any other purpose.
-It needs a little more resources to run than Nagios, but its footprint is still small.
-When the agent is installed on the monitored machines, an additional tool is installed too, which is used to send data to the Zabbix server asynchronously, similar to SNMP traps. Any program or script running on the machine can use this tool to send data to the server, which adds a great deal of flexibility.
-It is under active development, and although it is not as popular as Nagios, there are many people using it.

Disadvantages:
-The web interface is not very attractive.
-There is no native support for authentication when sending e-mail alerts. One can certainly make it work by installing and configuring some additional tools, but it requires some work.
-An agent needs to be installed on (most of) the monitored devices (having an agent could be an advantage in some circumstances).
-There is a network map option, but maps need to be created manually.
-There is an inventory option, but device details must be added manually.
-The default reports included are very few.
-The documentation isn’t very god.
-The install guide could be improved. The application needs to be installed from source, and resolving all dependencies to successfully compile the tool can take some time.

ZENOSS:

Advantages:
-Very nice web interface
-It performs agent-less monitoring. It uses WMI, SSH and SMTP to gather information from the monitored devices.
-Built-in support for authentication for e-mail alerts.
-Network mapping capabilities.
-Many reports included
-Backup and restore options from the web interface.
-Configurable dashboard
-It is under active development, and although it is not as popular as Nagios, there are many people using it.
-Acceptable documentation.
-Easy installation, with binary/native stack installers for most major Linux distributions. Some virtual appliances available as well.
-It can run Nagios plug-ins

Disadvantages:
-It is very resource-hungry when compared to Nagios and Zabbix.
-The web interfaces can be slow.
-It has device discovery capabilities, but devices need to be classified manually after the discovery, and monitoring starts after a process called ‘model’ has been run on the devices.
-Some information is stored in a database (events) but some other information is stored in text files.

This is a summarized comparison of the 3 options we evaluated. However it was not an in-depth evaluation, and there arem for sure, many things that we overlooked when making our choice.

Hope this helps someone in the future.

-CarlosDL

Thanks.

As I mentioned from the beginning, we have already tested Spiceworks.

http://www.spiceworks.com/referrer/mitrum

We use Nagios for all of our monitoring, mostly because of the power and flexibility of the system. We do not pay for anything Nagios, we use the service entirely free and have found NO LIMIT on the number or type of devices. We run Nagios on a Linux server, which works well. Despite it’s huge functionality, the footprint is actually quite small.

We use the Nagios OSS for free – it is a web-based service hosted on the server which runs the Nagios service. There are tools built in to allow for remote access and administration, so I can literally monitor every device on my network from anywhere in the world. The interface is very informational, and I do not find anything more than the web-based version to be necessary.

We do not pay for support. Nagios has such a large user base that I don’t find the support necessary. If you look around enough in forums and sites like this, you can find the configuration set ups and even code to do most everything you want.

The nice thing about Nagios is if you can think it, it can most likely be done. From monitoring temperatures, to printer jams, ink levels, camera packet loss, etc. basically everything can be done, and it supports Perl scripts…which expands its horizons even more! The extensibility of Nagios is phenomenal. It can get quite involved.

We have looked at Spiceworks, but mainly only to add the I.T. ticketing ability. It has not surpassed Nagios in any monitoring function.

Nagios can and will monitor disk usage, CPU load, processes or services, etc. It also has incredibly modifiable alerts (email 8am-5pm, text 5pm-9pm, off 9pm-8am, for example), if you’re willing to get your hands a little dirty.

Hope this helps!

-Schmidtw