[rrd-users] RRDtool spread values by time on rrd, how to turn off this behavior?

[Tobias Oetiker]

Today Daniel Hilst wrote:

> On 04/18/2012 10:35 AM, Tobias Oetiker wrote:
> > Hi Daniel,
> >
> > Today Daniel Hilst wrote:
> >
> > > Hey tobias, I'm using collectd tail plugin to monitor Oracle database
> > > logs,
> > > matching ORA- errors. When I got one error at 1334578215 for example, on
> > > graph
> > > I'll see 0.5 error at 1334578210 and 0.5 error at 1334578220 -- Remember
> > > that
> > > collectd uses 10 as step, so while generating graph I just multiply the
> > > value
> > > collected by 10. -- 0.5 error is not something readable, so I round epoch
> > > on
> > > rrdtool plugin to get an integer number, avoiding the interpolation.
> >
> > you may want to setup an RRD with 'ABSOLUTE' consolidation, then
> > rrdtool will store the error 'rate' which is what you probably want
> > ...
> >
> > cheers
> > tobi
> >
> Hey tobias, Thanks again for your reply,
>
> The weird thing is that, this time, I'm not trying to see a rate, but how many
> and when this errors ocurr. I'm using cacti with a lot of other graphs. On
> this ORA- errors graph I can see peaks when something bad happens. Using zoom
> feature from I can get peak hour and minute and see how many erros happen. Any
> way, ABSOLUTE consolidation is new for me -- rrdtool as a whole is -- I though
> thay was only AVERAGE, MAX, MIN and LAST ones. I'll study  ABSOLUTE with care,
> it may be what I want

I think so ... depending on the resolution of your RRD, the 'rate'
will be very high, given a sufficient zoom level ... obviously,
using RRD to store such information may not be ideal unless the
errors are more of a statistical thing than something which you
would want to track and resolve ... rrdtool is not ideal for sparse
data ...

cheers
tobi


>
>
> []'s
>
>

-- 
Tobi Oetiker, OETIKER+PARTNER AG, Aarweg 15 CH-4600 Olten, Switzerland
http://it.oetiker.ch tobi at oetiker.ch ++41 62 775 9902 / sb: -9900