[Icinga-devel] Icinga vs Nagios code structure

Hendrik Baecker andurin at process-zero.de
Thu May 7 22:13:18 CEST 2009

Hash: SHA1

Roy Sigurd Karlsbakk schrieb:
> Hi all
> Back in the Nagios 1.x days, I remember asking if it was hard/possible  
> to separate the nagios core with checks and so on from the web GUI,  
> since hardcoded HTML in C is rather ugly. I think this was just before  
> 2.0 was released and someone said the separation of the two was one of  
> the things that would make 3.x better. Now, that doesn't seem to have  
> happened. Nagios works well as it is, but it's a PITA to change the  
> web frontend.
The C coded CGIs are hard to customize, about multiple cascaded html
tables we are able to discuss if they are good or bad, fast or slow.
The way you need as a user to see the first hint of a problem and go
over to the details of that problem is very short, what is good in my

> Will Icinga try to change this? 
100% Yes. As I told Taylor in the other post (a few minutes back) we are
designing a new Webinterface unbound from the data source and a set of
normalized functions between them to be more modular.

> It'd be nice to have this in a client/
> server model so that the frontend could be written in some higher- 
> level-language like PHP while the backend remains as it is. This could  
> also allow for separation of the two if wanted/needed, and perhaps  
> with a distributed backend in case of large systems.

Are you talking about a thin client on your workstations or are we just
talking about using a webbrowser?
A real client software is a nice idea and might be possible in the
future if there is some SOA interface to the datasource. But we are far
far away from this right now, even if we talked already about things
like standarized web interfaces like SOA, XML, etc.
When we are thinking about this, it would be so cute to reuse the API on
the Serverside to just output some XML stuff, wouldn't it?

- -
Version: GnuPG v1.4.9 (GNU/Linux)


More information about the icinga-devel mailing list