I'd say what you are seeing is what is expected. For example if you take a
day worth of data 86400 sec at 15 sec resolution will give you 5760 daily
data points. If you are drawing a daily graph that is e.g. 500 x 300 you
will have to somehow consolidate all those data points. Ganglia uses
Averaging for consolidation. You may be able to use MAX instead but I
haven't actually tested it. Try adjusting the RRAs to use MAX instead of
AVERAGE.
Hope that makes sense.
Vladimir
On Wed, 13 Jun 2012, Alexander Karner wrote:
This is my RRA setup in gmetad.conf:
RRAs "RRA:AVERAGE:0.5:1:172800" \
"RRA:AVERAGE:0.5:4:525600" \
"RRA:AVERAGE:0.5:20:210240" \
"RRA:AVERAGE:0.5:40:210240"
--> I think I collect very detailed data for 172800 intervalls with 15s each.
This would equal 30 days.
As far as I understood the rrdtool docs there should be no average being
calculated for all data with this timeframe. (Just 1 piece of
data should be used for each datapoint)
But if I take a look into the Ganglia data I see some kind of mismatch with that theory.
Usually the "Max" values in the 1h, 2h and 4h
graphs are (mostly) the same (talking about a maximum within the last 60 minutes) but if
I look at the 1d graph the "Max" value as well
as the graph itself show smaller values. There is an additonal decreasement for
the 1week and 1month graphs (and values).
Is there a way to make better use of our RRA setup?
k
------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
Ganglia-general mailing list
Ganglia-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ganglia-general
