Ken - sorry for the late reply. I've been on vacation and just back now.
There are two concepts in monitoring. One is to "push" the information from a monitored system, the other is to "pull" it periodically from a central server. After years of working with multiple system types, from mainframes to micro's on various OS's, I've found the best approach is to have each node collect it's own information and then sweep by with the main system to collect it. This allows for the greatest resilience, scale, independence and so on. However, you're more than welcome to modify the system to your liking. My approach certainly isn't the only way to do it.
If you do change to another paradigm, make sure you blog it or comment it here - that allows others to learn from what you're doing.
Thanks!
- Buck
