You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
As a note, I think it is fine that we are only handling float64 histograms here, because
this code is just handling testcases.
Sum is the only value used here that can be int or float in metricdata.Histogram[N] (we don't read Min or Max) and Sum() always returns a float64, so there shouldn't be any runtime issues with this.
We force float64 when we handle Sum anyway, which is consistent with the behavior prior to this update and therefore not breaking.
However, we should follow this up with support for int64 histograms. Otherwise, users who switch to the GCP exporter might see slightly different values since we force floats in the output calculation. That would include adding a test case for an int64 histogram, which we don't appear to have currently. Will make a separate issue to track this.
Sum
is the only value used here that can be int or float inmetricdata.Histogram[N]
(we don't read Min or Max) andSum()
always returns a float64, so there shouldn't be any runtime issues with this.float64
when we handle Sum anyway, which is consistent with the behavior prior to this update and therefore not breaking.However, we should follow this up with support for
int64
histograms. Otherwise, users who switch to the GCP exporter might see slightly different values since we force floats in the output calculation. That would include adding a test case for an int64 histogram, which we don't appear to have currently. Will make a separate issue to track this.Originally posted by @damemi in #620 (comment)
The text was updated successfully, but these errors were encountered: