Let’s review following common situation – your company grow as well as amount of servers and applications in your web farm (yes, you still use own web farm instead of moder Cloud because of native “hard” code of your applications )
Of course you want to have a way to at least monitor servers and applications from centralized console!
More so, sometimes you also want to be notified by email or SMS that something going wrong with some of servers / applications.
How to made this wish real?

Solutions:

1. Write own low level code that will read and collect some way information from servers / applications. This way usually “hard” as you will probably spend hours to figure out how you can get CPU temperature, fan speed or how to made network ping in your favorite language using some low level OS API!
2. Use one of existed libraries and write application that can read and collect some way information from servers / application via this library. This way is much better and simple if compare to previous one! Library give you abstraction from low level system details (and it’s more important if you use different hardware / software architectures in your web farm). One of examples of such libraries is SIGAR (http://sourceforge.net/projects/sigar, http://support.hyperic.com/display/SIGAR/Home) – completely free and open source with support of amazing amount of OS / architectures). Another useful library – RRDtool -
   use it for data logging and graphing system for time series data and “to write your custom monitoring shell scripts or create whole applications using its Perl, Python, Ruby, TCL or PHP bindings”
   
3. Use one of already existed monitoring applications – Open Source / free or commercial. If you want open-source and free take a look into Nagios for example, with a lot of  plugins / addons (actually one way somewhere between 2) and 3) is to just create your own Nagios plugin and use it together with Nagios to monitor your servers and applications). But be ready that you will need to dial with a lot of installation / configurations issues and that you boss will not understand how he can use this soft (at least before you will found or build cool visualizer for Nagios). So in case if you did not dial with it before or you just want something “reach”, with nice UI etc, I recommend to give a try to ManageEngine Application Manager with ability to monitor up to 25 servers / applications in free version (in terms of ManageEngine you get “25 monitors” for free) or to try Hyperic HQ (exists in both Open Source and Enterprise versions). Both have nice UI, very simple installation (just currently Hyperic HQ needs to be installed manually for example on 64 bit Windows OS as no installation package exists for now for 64 bit  etc) and configuration. Both can be extended and adjusted very well for your situation. So review both applications and select one that feet better your needs (they have little different features and final cost will be different based on your situation. Generally as more servers / applications you need to monitor as more cheaper Hyperic will be if you will compare it to ManageEngine). Another well know solution is “Zenoss“, which have both commercial (really expensive – starting from 100$ per year per resource) and open source versions (sure open source have just “core” functionality, but it’s reach enough to take a look into it)

Maybe you know another way / library or application ? Let me know and I will update this post!