[SIG-scale] Meeting Notes 2025-10-09

[Andrew Burden]

10/9/25

Recording: https://youtu.be/f7JalC44u4Q 

Attendees: rhallisey, Sreeja Varnam IN

Agenda:

• Kubernetes 1.34 in performance job

• https://kubevirt.io/ci-performance-benchmarks/weekly/periodic-kubevirt-e2e-k8s-1.31-sig-performance/vmi/index.html

• https://github.com/kubevirt/project-infra/pull/4416

• Follow up PR to update the github page link

• Move env var from https://github.com/kubevirt/project-infra/blob/97d2a8cc5dcfad6aa3e6ec42f64730d06ff72f35/robots/cmd/perf-report-creator/scrape.sh to https://github.com/kubevirt/project-infra/blob/97d2a8cc5dcfad6aa3e6ec42f64730d06ff72f35/github/ci/prow-deploy/files/jobs/kubevirt/project-infra/project-infra-periodics.yaml#L1111 

• AIs from last time

• https://github.com/kubevirt/kubevirt/issues/15700

• Move KWOK jobs to 1.34 provider

• Increase KWOK scale from 1k vms to 4k vms

• Additional kwok metrics to gather:

• virt-controller reconcile queue metrics

• 4hr timeout

• https://github.com/kubevirt/project-infra/issues/4428

• Make cluster-clean error

• https://github.com/kubevirt/kubevirt/blob/main/hack/cluster-clean.sh

• Script errors when it doesn’t find and is unable to clean up a VMI

• Replace Client side rate limiting with APF

• https://github.com/kubevirt/kubevirt/issues/6307

• https://github.com/kubevirt/kubevirt/pull/15712

• CI now has CPU manager enabled, we should experiment with setting dedicated CPUs for the job

• Review cumulative for the last 2 weeks

• https://testgrid.k8s.io/kubevirt-periodics#periodic-kubevirt-e2e-k8s-1.34-sig-performance&width=20

• Depends on github page link