Bryan Boreham

~And more recently the [documentation and tests were added](https://github.com/prometheus/prometheus/pull/14849). Do we still need this PR?~ (This was irrelevant)

Hello from the bug-scrub! @bwplotka you may have most context on this - should we change the docs or change the behaviour? (In the first case we would close this...

I think the value is `false` to begin with, so there’s really only one check that sets it and the other one says to stop looking. It seems right to...

I couldn’t find a test case which broke with the previous code. However, traversing the path from the top of the expression is “obviously wrong”, and hence required some fix....

Can we update the proposed design in view of deciding to do histograms a different way? I see that `AppendCTZeroSample` was added in a different pattern. Can we make things...

Some previous discussion on CNCF Slack: https://cloud-native.slack.com/archives/C167KFM6C/p1710179010255979

I have rebased and updated the benchmarks. The regression seems to have gone away. I can no longer apply the same optimisation to `Reader` methods, because the indirection added in...

I ran the TSDB `PostingsForMatchers` benchmarks, and see what looks like a large percentage slow-down in many cases. However only the ones that take a couple of hundred nanoseconds slow...

Making it v2 would mean you can link both versions of the library, so we would also need to rename the metrics so both sets can be collected. My proposal...

Little bit of broader context: delayed name removal did break Mimir query sharding. https://github.com/grafana/mimir/pull/9577 I think this shows that the feature flag is safer.