[Icinga-devel] How much redesign, and when?
morten.guldager at gmail.com
Fri May 8 20:50:34 CEST 2009
Being a heavy nagios user, I have some sugestions for your fork. Even though
I could not figure out how fast you are planning redesigning "what works in
nagios". Anyway here comes my top-nags:
- need a more modular design (this is not a goal it self, more a way to make
other changes easier)
- move away from the hardcoded host/service concept. I quite often af hosts
without services at all, and I do also have cases where I need 3 levels (eg.
vmware: base, virtual machine, service), Instead have ONE object and make it
possible to have childs and parents.
- move the internal object database out to an module, make it possible to
chose diffrent storage engines. Personaly I would pick a mysql, but others
might continue with the files we know from nagios. This could lead us to the
ability to create new objects without restart.
- make it possible to implement mass-checker methods which could check
thousands of objects in parallel. Think of fping here. The forking-frenzy we
know from nagios is crazy. And implementing embeded perl code is dangerous,
its easy to kill the engine by mistake. Some sort of IPC mechanism could do
Well, thats my list, please share your thoughts, am I way out in the dessert
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the icinga-devel