Björn Rabenstein
Björn Rabenstein
> I'm confused here, cause these two expressions seem to have the same structure. the first one has better accuracy, but their performance shouldn't differ too much, CMIIW. `avg_over_time((min_over_time(some_metric{label="value"}[5m]) >...
Personally, I'm still not convinced adding `xrate` and friends to standard PromQL is a net gain for the community. However, coming from a very different angle, I am pretty convinced...
:facepalm: I think you are right @fpetkovski . Sorry @carrieedwards for not spotting that earlier. Maybe we can still utilize the `AllReverseBucketIterator` from this PR to make `histogram_quantile(1, metric)` more...
And we should add to the documentation of `histogram_quantile` how you can use the function to calculate an estimate of the highest/lowest observation in the range.
@carrieedwards just found that this is still open. Would you like to follow up on this as proposed above, i.e. amend the documentation of `histogram_quantile` to explain how to use...
Great, thank you very much. Please ping me on the new PR to avoid any delays.
@jesusvazquez could you pick this up?
@codesome what are the chances you can look at this any time soon? Otherwise, @jesusvazquez , would you be able to continue the review?
I'll review ASAP. But I'm afraid we need more expertise here. When I added the TODO about not supporting `AtHistogram`, I wasn't sure if that's the right way to go....
On the code level, this PR is OK once the open comments are addressed, but I would like to get a high-level confirmation from @codesome that it is OK to...