Re: [Shinken-devel] Questions about scheduler-broker communications, and livestatus module

[Laurent G. <lau...@ad...>]

Le jeudi 06 janvier 2011 à 15:50 +0100, nap a écrit :


> I'm just wondering if the "logs" are the best way for storing such
> reporting data. It's true it's already agregated, because we log the
> start of a problem, and the end too. So it's a good thing. For
> availability it sounds ok, and it's the major thing for such a
> reporting tool after all. So maybe yes, logs can be the data we need
> in fact :)

Right, nagios logs (status changes...) are actually the only way to keep
historical info for generating reports. They are precious and so must be
kept and be easily accessible, wherever they are stored (files,
databases...). And every state change message is precious too.

Standard nagios reports (availability, trends..) are based on nagios
logs, accessed directly with a standard Nagios, or via mklivestatus with
Thruk/Multisite. Also, many of my customers made custom scripts using
livestatus to access these historical data to make custom reports.

After that, it is always possible to translate these logs in a more
understandable format, why not.

Laurent