The one I'm writing will, first off, log all notifications to a file. Then that file will be periodically processed (via a cron-job) and notifications will be sent out by it. However, because it's batch-processing them, it can do nifty things like NOT SEND A SEPARATE PAGE FOR EVERY BLOODY PROBLEM. When you're monitoring over 1000 services, this is a nice thing.
My goal is to make a robust but sensible paging mechanism, designed to work well for those who are monitoring large-scale environments and simply can't afford to get paged for every single problem that occurs. I'm still trying to decide how exactly I want to manage it.
I may make it an entirely separate 'product' to add-on to Netsaint, as I might end up having to replicate the entire notification scheme. I'd like to work _with_ Netsaint, at least to handle things like the time periods and such. I don't want my program to have to dig through the Netsaint config files in addition to its own and do all of that work. I also want Netsaint to handle the other timing - i.e. re notifications and the like.
So, anyway, that's my goal. It definitely needs reworking, I just need to figure out the best way.
Anyway, now that my pager has been quiet for 20 minutes straight (a record for tonight :/ ), I'm going to assume it's napping and try to get some sleep myself.