This also would be inaccurate.
The method we use generates an accurate figure for graphs using 5 min samples, beyond that it’s not possible.
If one port does 999mb at the start of a 30 min period and another does 999mb 25 mins later, we’d generate an aggregate max of 1998mb n any graph with these two ports and once they are averaged to 30min samples, which is obviously useless :)
There’s no actual solution to this, the data simply no longer exists at this point. We should /probably/ hide the aggregate max on graphs > 24hr
Adam.
Sent from my iPhone
On 6 Oct 2022, at 03:57, David Milton via observium observium@lists.observium.org wrote:
Thanks for the explanation. To my mind however, the aggregate max would be to take the sum of all values of the same time and select the maximum sum. Using an average for a maximum when all other values are discrete produces results that aren’t useful. I know it’s not Observium but that’s a major problem with rrdtool in my opinion
On Wed, 5 Oct 2022, at 13:08, Adam Armstrong via observium wrote: Hi David,
This is due to how rrdtool works. It doesn't make a lot of sense to aggregate the max of max together, so the max for the aggregate is the max of the aggregated averages, rather than the max of the max values, which wouldn't make a lot of senses since it'd be merging peaks which may not have actually occurred at the same time.
adam.
David Milton via observium wrote on 05/10/2022 17:55:
Hello,
We have been seeing some results in our stacked graphs that do not appear to make sense. Initially we thought this was because the maximum values for the two values occurred at different points in time. That cannot be the whole issue because even if you take the sum of every pair of numbers the maximum aggregate must at least equal the maximum value of one of the two series.
This graph shows the problem:
How can we have a maximum value of 97.4M but an aggregate maximum of 55.7M? It seems that maximum aggregate should also at least be 97.4M.
What’s happening here?
Thanks, Dave.
observium mailing list -- observium@lists.observium.org To unsubscribe send an email to observium-leave@lists.observium.org
-- Sent from Postbox https://www.postbox-inc.com/?utm_source=email&utm_medium=siglink&utm_campaign=reach _______________________________________________ observium mailing list -- observium@lists.observium.org To unsubscribe send an email to observium-leave@lists.observium.org
-- David Milton david@dmilton.ca, For better email, sign up here: http://www.fastmail.fm/?STKI=7947829 _______________________________________________ observium mailing list -- observium@lists.observium.org To unsubscribe send an email to observium-leave@lists.observium.org