Today, Steven Wagner wrote forth saying... > > i'm going to update gmetad soon (today?) to incorporate some of the new > > enhancements. at the minimum i'll have gmetad skip any metrics with a > > SLOPE of ZERO. that'll save some serious disk I/O. i haven't tested it > > but i'm 90%+ certain that gmetad is fine with the new changes (it just > > will ignore them). > > I don't know if RRDs like it when you skip data points like this. In fact > I am pretty sure they don't.
what i meant by ignoring is that the old gmetad will not even process the TN and TMAX attributes.. it'll just ignore them. the new gmetad will however use the TMAX and SLOPE when creating the rrdb in the first place. that'll allow less updating of some rrdbs who's metric has coarse time granularity. -matt
