[rrd-users] Re: MRTG and RRDTool...
cliftonr at lava.net
Wed Nov 14 21:12:05 MET 2001
On Wed, Nov 14, 2001 at 11:40:18AM -0800, Lightfoot, Kirk wrote:
> Is it possible to use MRTG as my frontend, and have it feed data into an RRD
> that only stores 5 minute samples and never averages them into 30-minute and
> 2-hour samples?
> I'd like to have an RRD that stores and graphs based on samples that are fed
> every 5 minutes from MTRG, and save up to 1054080 of them? (approx 10 years
> of 5 minute samples)
> I don't want any diluted graphs based on 30-minute or 2-hour averages.
So are you going to draw your ten-year graphs 1,054,080 pixels wide?
I understand wanting to keep high precision, but it seems to me
you've got to do some averaging at some point to have usable graphs.
Thinking about what you'd like the intervals for those per-pixel
averages to be, for the one and ten year graphs, probably would tell
you where it makes sense to do averaging in your RRD.
There might be some statistical purposes where it would make sense to
store all the individual 5-minute samples, but for typical purposes
including graphing, it would seem to me that the consolidation
functions would be much more useful. (It will also avoid incurring
roughly 8MB of storage - 1,054,080 IEEE double values per data source -
for every data source you track, which could stack up pretty fast.)
RRD can do it if you want to, and I'm sure you can modify MRTG's
defaults to create it that way; I'm just advising caution.
Clifton Royston -- LavaNet Systems Architect -- cliftonr at lava.net
WWJD? "JWRTFM!" - Scott Dorsey (kludge) "JWG" - Eddie Aikau
Unsubscribe mailto:rrd-users-request at list.ee.ethz.ch?subject=unsubscribe
Help mailto:rrd-users-request at list.ee.ethz.ch?subject=help
More information about the rrd-users