[rrd-users] Re: MAXing or AVERAGing 1 steps

[Tobias Oetiker]

Yesterday you sent me mail regarding [rrd-users] MAXing or AVERAGing 1 steps:

*> 
*> I wanted to keep a log of exactly how many modems are in-use, with
*> one-minute samples.  I thought I'd put it into and RRD to make it easy to
*> graph (instead of using gnuplot or something).  I made a simple rrdtool
*> source:

if you updates arrived EXACTLY at the points in thime rrdtool
stores data in its database (963600900,963600960,963601020,...)
then exactly these values would be stored ... but because your data
does not arrive exactly at these points in time rrdtool integrates
the space below the curve created by the datapoints you supplied
and then uses the --step size to calculate the value of the data
source at the points in time which havbe to go to the log file.

An example ... you know tat at 8:01am there were 5 modems in use.
At 8:07am there were 8 modems in use. If you now tell rrdtool to
log data every 5 minutes it will log data 8:05 ...
what should it log ? soemthing in between 5 and 8 ... 

that is what you see in the log file ... this is fine ... it is an
average ... no need to worry ... this is exactly what you want ...
MAX at a resolution of one interval does not make sense ...

make sure you work through the tutoral ... it will help you
understand theree things better ....


cheers
tobi

*>  create foo.rrd --step 60 \
*>  DS:modems:GAUGE:600:0:24 \
*>  RRA:AVERAGE:0.5:1:3600 \
*>  RRA:MAX:0.5:1:5760 \
*> 
*> When I dump the data with "AVERAGE", ...I get:
*> 
*>  963600900:         15.25
*>  963600960:         15.43
*>  963601020:         14.32
*>  963601080:         15.35
*>  963601140:         15.62
*> 
*> There were probably not 15.62 modems in-use at 963601140.  
*> When I fetch the data with "MAX", I get:
*> 
*>  963612780:         14.00
*>  963612840:         14.00
*>  963612900:         14.83
*>  963612960:         15.00
*>  963613020:         15.00
*>  963613080:         15.00
*> 
*> Similarly probably not 14.83 modems in-use at 963612900.
*> 
*> I thought that the "RRA:AVERAGE:xff:1:whatever" would cause "update" to
*> store an "average" of a group of 1 samples (therefore no averaging at
*> all), and the data in the log would match the "update" values exactly; and
*> similarly for MAX:xff:1:whatever .
*> 
*> I can certainly just store the values in a log-file for such a simple
*> purpose; but it would be good to know what's going on here, since we work
*> on the assumption that the AVERAGE:xff:1 RRA contains the unchanged values
*> that were stored with "update".
*> 
*> -Bradley
*> 
*> 
*> --
*> Unsubscribe mailto:rrd-users-request at list.ee.ethz.ch?subject=unsubscribe
*> Help        mailto:rrd-users-request at list.ee.ethz.ch?subject=help
*> Archive     http://www.ee.ethz.ch/~slist/rrd-users
*> WebAdmin    http://www.ee.ethz.ch/~slist/lsg2.cgi
*> 
*> 

-- 
 ______    __   _
/_  __/_  / /  (_) Oetiker, Timelord & SysMgr @ EE-Dept ETH-Zurich
 / // _ \/ _ \/ / TEL: +41(0)1-6325286  FAX:...1517  ICQ: 10419518 
/_/ \.__/_.__/_/ oetiker at ee.ethz.ch http://ee-staff.ethz.ch/~oetiker


--
Unsubscribe mailto:rrd-users-request at list.ee.ethz.ch?subject=unsubscribe
Help        mailto:rrd-users-request at list.ee.ethz.ch?subject=help
Archive     http://www.ee.ethz.ch/~slist/rrd-users
WebAdmin    http://www.ee.ethz.ch/~slist/lsg2.cgi