Nagios is the best software package I know of for monitoring small to medium size computer networks. I have absolutely no complaints with the software anymore. I find it very robust and it scales well as my network grows.

I do hear from others who seem to have Nagios problems from time to time while I peruse the forums (I suggest everyone check the forums that interest them to stay informed and up to date on new technology). The main complaint I hear about from some administrators is that they think Nagios emails them to often when the network has problems. There assumption that Nagios is emailing them to often is a wrong headed theory.

There actual problem is two fold. The first problem is they have too many actual errors on the network that they are not clearing and the second is that they don't have their Nagios configuration setup properly. The standard configuration of Nagios is setup to work with the broadest possible group of networks so it is a very basic setup.

The configuration files that need modifying to setup your notifications in a sane way are all in the etc directory. I like to setup contacts.cfg first. I will setup the IT group email accounts for 24 by 7 email notification. I then setup individual admin telephones for SMS and put them on 24 by 7 notifications. I will usually setup plenty of other notifications for managers and office personnel.

I like to have at least two contactgroups that I belong to. I setup an admin-email, admin-text-message in contactgroups that I belong to. I also like to have at least three different types of host the generic-host and the important-host and critical-host. The only difference between my host are there notification options and intervals. The less critical notifications have longer intervals and less options.

The real work comes in my services.cfg file. I keep the generic-service, but I also will setup an important-service and critical-service. The services on infrastructure devices like routers, switches, file servers, license server, computer rooom temperature monitor are critical to me, so there most important services are "use" defined as critical-service. Desktops that I label as generic-host have services that are labeled generic-service. Simulation server services are defined as important. I like to have those computers online all the time but I do have a lot of them so I am not going to come into the office at midnight because one of them goes down.

Once you use the configuration wisely you will receive many less nuisance contacts and the notifications you do receive you will know are for the services and host that matter the most to you.

When you receive notifications they are sent out as either host notifications or service notifications. I don’t receive notification of any type for generic-host or services. I notice that they have a problem when I look at my monitoring servers Nagios web page. Important host and service notifications are sent to my email so I know they have a problem as long as I am able to get email. I get a text message when a critical-host or critical-service has a problem.