[rrd-users] RRDTool Aggregation Inaccuracies.

[Carson Gaspar]

NHarris wrote:

> Second, I assumed that the aggregation (via CDEF) was done and THEN a
> maxima was pulled from the new aggregation

If you _really_ need the true maxima for (A+B), you should be able to create a 
COMPUTE dataset to do so. But I've never played with COMPUTE datasets, so my 
knowledge is just from the man page.

You can fake it by graphing MAX(AVG(A) + AVG(B)), but you won't get true maxima. 
  It will be fairly close, though.

-- 
Carson