Wojciech Tyczynski

/remove-lifecycle stale /lifecycle frozen

/remove-lifecycle stale

/remove-lifecycle stale /lifecycle frozen

/remove-lifecycle stale

/remove-lifecycle stale

I've seen a huge improvement in network programming latency for our scalability tests: https://perf-dash.k8s.io/#/?jobname=gce-5000Nodes&metriccategoryname=Network&metricname=Load_NetworkProgrammingLatency&Metric=NetworkProgrammingLatency And I initially thought that it would be this PR, but the timing doesn't match -...

heh, it seems I've responded myself. The FG was originally disabled in our tests, and it god enabled by this PR: https://github.com/kubernetes/test-infra/pull/27927 The timing matches. So the summary is -...

> I don't know exactly what the perf tests do, but the improvement should be greatest in the case where there are a steady stream of small changes, which kube-proxy...

> I have a bit more started in my local branch (https://github.com/kubernetes/kubernetes/pull/113666 is a preparation for making that possible), but my work (for now) is limited to lists. and watches

/remove-lifecycle stale