Nagios used for server energy consumption monitoring


I intend to use Nagios in a project related to monitoring the energy consumption of a server. Since I have little experience with network monitoring tools and I find the Nagios documentation to be somehow disorganized I would appreciate some help on where to start my research. So, I have a few question:

  1. What is the performance impact of Nagios on the system where it is installed? Basically I’m interested if there is a significant performance penalty in installing Nagios locally on each machine vs monitoring the machine remotely from a central Nagios server?
  2. What is the architecture of Nagios (the basic components, their function and their interaction with the host) and which are the recommended deployment scenarios?
  3. When it comes to monitoring energy consumption what are my alternatives (plugins and devices that integrate with Nagios which can measure energy consumption, plugins that identify the models of various system components)?
  4. I am thinking about launching an external application written in Java when there is a constant overflow of a certain threshold (e.g. CPU load over a certain limit for some extent of time, occupied system memory reaches a certain critical limit etc.). Would you recommend using the NEB in order to hook up to certain events (also, could you point to the new NEB documentation for Nagios 3.x), or periodically polling Nagios itself from my application in order to retrieve the necessary information.