[rrd-users] Graphing average temperatures
Erik de Mare
erik at oezie.org
Thu Nov 22 08:40:08 CET 2007
I use RRA's to calculate the averages.
my stepsize is 5 min(300sec) with a script I read it from a DS1820 maxim IC.
I have the following RRA's defined.
RRA:AVERAGE:0.5:1:766 <<- 2,5 days of 5 min samples.
RRA:AVERAGE:0.5:6:766 <<- 2 weeks data with the average of 30 mins(6*300sec)
RRA:AVERAGE:0.5:24:766 <<- 2 months data with average of 2 hours(24*300sec)
RRA:AVERAGE:0.5:288:766 <<- 2 years of data with the average of 1 day(288*300sec)
I also have these for MIN and MAX. when I draw a year graph the last RRA will be used by rrdtool. but you can also graph 2 months with the last RRA, then you have to 'force' rrdtool to use that specific rra in the DEF statemen, example DEF:..:..:..:..:step=86400
the sliding average is just a CDEF with TREND.
I hope this makes it a bit more clear to you.
On Thu, 22 Nov 2007 07:40:06 +0200, "Petteri Matilainen" <pmatil at gmail.com> wrote:
> Yes, I think the sliding window average (TREND) could work for me. So I
> to define a 24 hr sliding window to plot 24 hr averages? Is the TREND
> same as if I graphed this: GPRINT:Inside:AVERAGE:"\t%6.2lf °C" (on the
> picture on my last post)? Erik De Mare: how do you graph the averages for
> each day?
> On 11/22/07, Erik de Mare <erik at oezie.org> wrote:
>> I think you are looking vor something like this:
>> the thick lines are TREND lines with a window of 1 week.
>> the thinner lines are the average values for that day, and the most thin
>> lines are min and max values during that day.
>> Petteri Matilainen schreef:
>> > If I'm reading your graphs correctly, you don't graph the average
>> > temperature at all. You only display it on the 'avg.' column. I
>> > specifically need to graph the average over 24 hrs, 1 week, 1 month
>> > 1 year. I have read your 'how to do this' and it has been very useful.
>> > Here's a sample image of what I've managed to graph so far:
>> > http://www.kotikone.fi/pmatil/out_day.png.
>> > On the picture there is also the gap I mentioned.
>> > The average curve is plotted the way I described it on my earlier
>> > cheers,
>> > Petteri
>> > On 11/21/07, *Paul Rimmer* <paul-rimmer at ronin-tech.com
More information about the rrd-users