Bryan Boreham
Bryan Boreham
I think it's more like #13643, which was a bad merge, fixed by #13702. There is a release candidate for 2.51.0 out now which has this fix.
We need to wait longer than 1 day for feedback to come in on the release candidate. Please try it out! If nobody has any feedback this then will become...
There is already a bucketed pool [here](https://github.com/prometheus/prometheus/blob/c954cd9d1d4e3530be2939d39d8633c38b70913f/util/pool/pool.go#L22-L24), used in scraping. Note I did some experiments with that pool, including taking out the `interface{}` at #12972. If you do decide to...
From Prombench this PR is doing more allocations with a bigger heap, however CPU is lower. This is confusing. If it wasn't branched off latest main I suggest you rebase...
There's no value in guessing; we have tools such as profilers to look at the detail. However, it's a short PR: there is no mechanism by which it can reduce...
I haven't looked at the test at all. Since it's only flaking on Windows, could it be the timer resolution again?
I just got this on my desktop Linux VM: ``` $ go test -tags stringlabels ./rules/ --- FAIL: TestAsyncRuleEvaluation/asynchronous_evaluation_with_independent_and_dependent_rules (0.06s) manager_test.go:2391: Error Trace: /home/vagrant/src/github.com/grafana/mimir-prometheus/rules/manager_test.go:2391 Error: "0.056689165" is not less than...