[rrd-users] filter rrd data

[Maxim Kuleshov]

Hi!

I heavily use rrdtools in self-written monitoring facilities. While
debugging and fixing scripts, errors in collection data happen and
incorrect values (often much greater than average) being store in .rrd
files. Therefore, I see high peaks on graphs and usable info is scaled
down.

I know it is possible to set maximum for basic filtering, although often
incorrect numbers passes.

I could manually dump .rrd, edit xml and then restore it, but may be
there is a tool (third-party?) performing filtering raw values
(already stored in file) for specified boundary, substituting "bad"
values for "unknown"?