[rrd-users] Min Max Values

Steve Shipway s.shipway at auckland.ac.nz
Thu Mar 12 21:50:24 CET 2015


> I'm a bloody beginner.
> I stred some DS in a RRD. All works fine.
> I use
>   DS:i1:GAUGE:60:-10:100 \
>  RRA:AVERAGE:0.5:1:86400
> 
> I understand: over 60 sek, there will be a average value storeed in the
> database Thats fine, but I like to have the max und min values additional.
> All I found a description for data consollidation for min/max.

Your DS line defines the data source; this says that you will get a new data
sample every 60 seconds (the base sample interval), that can be between -10
and 100.

Your RRA specifies how this will be stored; in this case, you will keep
86400 consolidated points, where each one is an average of a set of just 1
sample.  In other words, you'll store this data point (after normalisation!)
and keep 86400 of them.  They will be spaced 60s apart (the number of points
in the set, multiplied by the sample interval), and so will cover one day.

While you can also define a MAX:0.5:1 and MIN:0.5:1 RRA, a little thought
will show that the maximum and minimum of a set of 1 sample will be the same
as the average of that set!  There is only a difference when you are
consolidating more than 1 in your set...

If you want to know the min/max over the 1min interval, you need to take
more samples.  For example, you could sample every 10 sec, and make your RRA
an AVERAGE:0.5:6 -- average of a set of 6.  This will make the RRA
consolidated data points 60sec apart as in your first example, but a
corresponding MAX:0.5:6 will likely be different...

(Note: the 0.5 in the RRA is the fraction of the consolidated data that is
permitted to be Unknown.  Generally, to always have this as 0.5 is fine)

Hope this makes it clearer.

Steve



Steve Shipway
s.shipway at auckland.ac.nz

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 10311 bytes
Desc: not available
URL: <http://lists.oetiker.ch/pipermail/rrd-users/attachments/20150312/8d60d704/attachment-0001.bin>


More information about the rrd-users mailing list