[collectd] Plugins values

Frédéric Pégé frederic.pege at gmail.com
Fri Feb 8 11:56:08 CET 2013


Hi,
and thanks for the answer.
As for the tendency, I would prefer the mutliple datasource way. I still
don't get why all the values have to end in ".value", and be in separate
"folders". But nevermind.
As long as it is done the same way in every plugin, it's more logical.

Can I help in any way possible ?
df, load, cpu, ... and other "standards" plugin are in C, right ?

Fred.


2013/2/8 Fabien Wernli <collectd at faxm0dem.org>

> Hi,
>
> On Thu, Feb 07, 2013 at 01:33:19PM +0100, Frédéric Pégé wrote:
> > The load values are returned in 3 different values under the same item.
> > load/load. Values are longterm, shortterm and midterm. And I'm ok with
> that.
>
> Just my two cents on loadavg: IMHO it's totally useless to keep anything
> else than shortterm, if you're using some kind of historization like
> rrdtool, as you'll get the two others for free.
> midterm/longterm are only useful when no history is available,
> e.g. from the command line, or if you're using some kind of realtime
> visualization.
>
> > Do you have an explanation for these different ways ?
>
> Anyways, back to your real question:
> The reason for the differences is purely historical. The tendency is to
> move
> away from multiple datasources (like load, if_octets, etc.) and use only
> simple datasources, like gauge, cpu, etc.
>
> Maybe this will happen in collectd 6, if two wisely chosen guys have the
> courage to do so ;-)
>
>
> _______________________________________________
> collectd mailing list
> collectd at verplant.org
> http://mailman.verplant.org/listinfo/collectd
>



-- 
Linux: Rejoins-nous, nous sommes déjà 11.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.verplant.org/pipermail/collectd/attachments/20130208/fb327367/attachment.html>


More information about the collectd mailing list