[SIG-scale] Meeting Notes 2026-08-06

[Andrew Burden]

8/6/26

Recording: https://youtu.be/ZXgSUEYHLoI 

Attendees: rhallisey

Agenda:

• [alaypatel07] 1.36 lanes have regressed in vmi creation to running by more than 80%

• https://storage.googleapis.com/kubevirt-prow/logs/periodic-kubevirt-e2e-k8s-1.36-sig-performance/2078027029990608896/artifacts/VMI-perf-audit-results.json 

• https://kubevirt.io/ci-performance-benchmarks/weekly/periodic-kubevirt-e2e-k8s-1.36-sig-performance/vmi/index.html 

• Breakdown phase transition times for vm creation

• Access the CI cluster and run the perf job locally

• Look for the root causes for the below scale/performance changes

• https://kubevirt.io/ci-performance-benchmarks/weekly/periodic-kubevirt-e2e-k8s-1.34-sig-performance/vmi/index.html 

• List Nodes count VMI

• Drops to 0 around Jun 7, 2026

• https://github.com/kubevirt/kubevirt/pull/16106

• Migrates from the kubevirt client to kubernetes for metrics

• We’re no longer gathering the Node list count

• Is there a way to pull these metrics from the k8s client?

• Beta flag performance - https://github.com/kubevirt/kubevirt/pull/17405

• Open a PR to move each Beta feature to alpha.

•   {"name": "ExternalNetResourceInjection", "state": "Beta"},

•   {"name": "ImageVolume", "state": "Beta"},

•   {"name": "KubevirtSeccompProfile", "state": "Beta"},

•   {"name": "LiveUpdateNADRef", "state": "Beta"},

•   {"name": "MigrationPriorityQueue", "state": "Beta"},

•   {"name": "NodeRestriction", "state": "Beta"},

•   {"name": "PasstBinding", "state": "Beta"},

•   {"name": "PodSecondaryInterfaceNamingUpgrade", "state": "Beta"},

•   {"name": "SecureExecution", "state": "Beta"},

•   {"name": "Snapshot", "state": "Beta"},

•   {"name": "VideoConfig", "state": "Beta"},

• Periodic Job run 7 days a week