Björn Rabenstein

[This issue was discussed at the last dev-summit.](https://docs.google.com/document/d/1uurQCi5iVufhYHGlBZ8mJMK_freDFKPG0iYBQqJ9fvA/edit?pli=1&tab=t.0#heading=h.oe6cijk5g79) To summarize: There were no objections, but some desire to have a more detailed discussion. @krajorama , @roidelapluie , and myself were...

My current thoughts: - As already discussed above, I retract my concern about variadic functions. They do exist in PromQL already. - What I believe is still a new concept...

Yes, let's put it under the already existing `--enable-feature=promql-experimental-functions` flag so that we can tweak details without cutting a major release.

Hello from the bug scrub! An open PR is under review for this, leaving open.

Nominating @bboreham , who has reviewed https://github.com/prometheus/prometheus/pull/11474 .

Thanks @aknuds1 . This makes sense at first glance. @wangzhao765 would you like to pull @aknuds1's commit into your branch? Alternatively, @aknuds1, you could just create a new PR from...

If you used their commits in your work, I would totally include them. Depending on the final commit structure, we might squash them all into one commit, but both of...

With no response from @wangzhao765 , I close this in lieu of https://github.com/prometheus/prometheus/pull/15719 (which is already under active review by @dgl ).

The summaries use the classical algorithm often called CKMS (by the initials of its authors), see https://grafana.com/blog/2022/03/01/how-summary-metrics-work-in-prometheus/ for some detail. The Go implementation was originally by bmizerany. One day, we...

FTR: We'll have a detailed look here once #16562 is merged.