[rrd-users] can RRDTool help me? merge datasets

p vr flexman75 at gmail.com
Wed Apr 18 09:20:39 CEST 2007


i am brand new to rrd and was wondering if rrd can help me with the
following situation, i have two time series:

1. concurrent users/time
2. connections/time

it is not obvious in this hypothetical example but in my situation i am not
able to sample my data from the same place, so each series is sampled
against it's own timeframe at roughly the same interval (2 seconds). now i
want graph connections vs users. according to the documentation rrdtool is
able to interpolate the datasets, so effectively i assume it should be able
to merge these datasets when i insert them into the same rrd. i created a
rrd like this:

rrdtool create user-rra-max-registerGroup.rrd --start 1176840612 --step 2
DS:user:GAUGE:15:0:10000 DS:connections:GAUGE:15:0:10000 RRA:AVERAGE:0.2

i insert each series of data independently with a perl script, basically

rrdtool update my.rrd -t user $timestamp:$users
rrdtool update my.rrd -t connections $timestamp:$connections

these two series' of data was captured at overlapping time, however after
the insert fetching from the db they appear to be mutually exclusive:

1176842132: 3.0200000000e+02 nan
1176842136: 3.0275000000e+02 nan
1176842140: 3.0300000000e+02 nan <--- users ends here
1176842144: nan 4.7055000000e+03 <--- connections starts here
1176842148: nan 8.1140000000e+03
1176842152: nan 3.4725000000e+03

even though the connections series started earlier in time.

am i misunderstanding what rrdtool can do?

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.oetiker.ch/pipermail/rrd-users/attachments/20070418/ec8f3777/attachment.html 

More information about the rrd-users mailing list