>>>> if you update more often than required for you step size, the data will >>>> be aggregated using the aggregation function you've chosen (MIN, MAX, >>>> AVERAGE etc) so either the smallest or the highest or the average of the >>>> values will be stored. >>> >>> No, the aggregation functions are NOT used for sub-step updates. >>> >>> See Alex's tutorial on rates and normalising - it's all explained well >>> there. >>> http://www.vandenbogaerdt.nl/rrdtool/ >> >> Actually I did not really describe sub-interval updates. And I don't know >> from the top of my head. I'd have to look in the sources to see what really >> happens when you do that. > > Well it's "sort of" in there, but not really highlighted. If you > update every second with a step size of 60, then you can feed in one > value of 60 and 59 values of 0 - and get the result of 1/second for > the step. Min. Max, etc only come in when consolidating PDPs. > > That's what I've observed anyway.
Agreed. That's what I've observed, too. Sorry for the somewhat misleading info in my first reply. I'm using mostly AVERAGE as aggregation function, so I got confused myself, but aggregation happens at the RRA-entry-level. For PDPs it seems to be always average. - Karl _______________________________________________ rrd-users mailing list [email protected] https://lists.oetiker.ch/cgi-bin/listinfo/rrd-users
