Blocking all CI payloads, failing on aws/azure and gcp multiple in a row. Sample payload: https://amd64.ocp.releases.ci.openshift.org/releasestream/4.20.0-0.ci/release/4.20.0-0.ci-2025-06-26-071316

looks like it is caused by https://github.com/openshift/cluster-kube-apiserver-operator/pull/1846 which Ben believes is surfacing a pre-existing problem more clearly. https://issues.redhat.com/browse/OCPBUGS-55626 Ben is pursuing a revert.

Hits tests which run in aggregated jobs:

kube-api-http1-service-network-reused-connections mean disruption should be less than historical plus five standard deviations
kube-api-http2-service-network-reused-connections mean disruption should be less than historical plus five standard deviations
openshift-api-http2-service-network-reused-connections mean disruption should be less than historical plus five standard deviations