briareos at eswat.org
Thu May 19 16:52:31 CEST 2011
Thanks Manuel for your reply,
my mail was meant to be sent to the list, but the reply-to field points to you,
so I'm forwarding your reply to the list right now.
>On Thu, May 19, 2011 at 2:44 PM, Mirko Buffoni <briareos at eswat.org> wrote:
> > At 13.13 19/05/2011 +0200, you wrote:
> >> One more thing: in the test2 you can use the mouse to move the graph:
> >> click and drag to move sideways, and to zoom use the wheel.
> > WOW!
> > First of all, my hat off for such a great achievement.
> > I let my jaw literally fall down after seeing 'test2'.
> > So this is a 'Bravo' for sure!
> > Now the only drawback I see in this design is the amount of data which
> > is transferred to the client. Basically you're parsing a real rrd database
> > transferred and cached by the client browser, and which basically
> > changes of a few bytes every tick, but that to the browser
> > appears to be refreshed entirely. This would be a hit for graphs which are
> > an aggregate of more .rrd files (dns/mysql to cite some).
> > An interesting point would be to ask the server to send back data for
> > the construction of the canvas graph. But how much useful would this be
> > over a plain .png generation? I think I'd have to check how google caches
> > data for graph generation in their finance webapp.
> > Anyway, congratulation again for your work! :)
> > Mirko
>The idea for the rrd files was to test the rrdgrah using the command
>line syntax of rrdtool,
>We can implement another function like rrd_fetch to download the data
>via a json like Collection4 do.
>Another idea could be update the graphs in real time without
>downloading all the data, or the data of last hour,
>I'm not sure about the way to do this with collectd, maybe with a
More information about the collectd