kops scalability tests doesnt have a way to configure scheduler QPS #17751
Description
/kind bug
1. What kops version are you running? The command kops version, will display
this information.
master branch
2. What Kubernetes version are you running? kubectl version will print the
version if a cluster is running or provide the Kubernetes version specified as
a kops flag.
master branch
3. What cloud provider are you using?
GCE
4. What commands did you run? What is the simplest way to reproduce this issue?
/tests/e2e/scenarios/scalability/run-test.sh
5. What happened after the commands executed?
This PR creates an env variable to allow for qps configuration for kube-scheduler #17742. However, it is not respected.
I1110 03:09:34.558041 12 configfile.go:94] "Using component config" config=<
apiVersion: kubescheduler.config.k8s.io/v1
clientConnection:
acceptContentTypes: ""
burst: 100
contentType: application/vnd.kubernetes.protobuf
kubeconfig: /var/lib/kube-scheduler/kubeconfig
qps: 50
The component config that is uses a lower qps burst value then the flags:
I1110 03:09:34.550578 12 flags.go:64] FLAG: --kube-api-burst="500"
I1110 03:09:34.550580 12 flags.go:64] FLAG: --kube-api-content-type="application/vnd.kubernetes.protobuf"
I1110 03:09:34.550582 12 flags.go:64] FLAG: --kube-api-qps="-1"
6. What did you expect to happen?
I expected the configuration from run-test.sh to change the component config so that scheduler is configured with right qps for the test.
7. Please provide your cluster manifest. Execute
kops get --name my.example.com -o yaml to display your cluster manifest.
You may want to remove your cluster name and other sensitive information.
apiVersion: kops.k8s.io/v1alpha2
kind: Cluster
metadata:
creationTimestamp: "2025-11-10T03:05:58Z"
name: scalability.k8s.local
spec:
api:
loadBalancer:
type: Public
authorization:
rbac: {}
channel: stable
cloudConfig:
gceServiceAccount: default
cloudControllerManager:
concurrentNodeSyncs: 10
cloudLabels:
group: sig-cluster-lifecycle
subproject: kops
cloudProvider: gce
configBase: gs://k8s-infra-e2e-scale-5k-project-state-adfe/scalability.k8s.local
containerd:
configAdditions:
plugins."io.containerd.grpc.v1.cri".containerd.runtimes.test-handler.runtime_type: io.containerd.runc.v2
etcdClusters:
- cpuRequest: 200m
etcdMembers:
- instanceGroup: control-plane-us-east1-b
name: b
volumeIops: 10000
volumeSize: 120
volumeThroughput: 1000
volumeType: hyperdisk-balanced
manager:
backupRetentionDays: 90
env:
- name: ETCD_QUOTA_BACKEND_BYTES
value: "8589934592"
- name: ETCD_ENABLE_PPROF
value: "true"
listenMetricsURLs:
- http://localhost:2382
memoryRequest: 100Mi
name: main
- cpuRequest: 100m
etcdMembers:
- instanceGroup: control-plane-us-east1-b
name: b
volumeIops: 10000
volumeSize: 120
volumeThroughput: 1000
volumeType: hyperdisk-balanced
manager:
backupRetentionDays: 90
env:
- name: ETCD_QUOTA_BACKEND_BYTES
value: "8589934592"
- name: ETCD_ENABLE_PPROF
value: "true"
memoryRequest: 100Mi
name: events
iam:
allowContainerRegistry: true
legacy: false
kubeAPIServer:
anonymousAuth: true
enableContentionProfiling: true
enableProfiling: true
logLevel: 2
maxRequestsInflight: 4000
kubeControllerManager:
enableContentionProfiling: true
enableProfiling: true
endpointSliceUpdatesBatchPeriod: 500ms
endpointUpdatesBatchPeriod: 500ms
kubeAPIBurst: 500
kubeAPIQPS: "500"
kubeProxy:
metricsBindAddress: 0.0.0.0:10249
proxyMode: nftables
kubeScheduler:
authorizationAlwaysAllowPaths:
- /healthz
- /livez
- /readyz
- /metrics
enableContentionProfiling: true
enableProfiling: true
kubeAPIBurst: 500
kubeAPIQPS: "-1"
kubelet:
anonymousAuth: false
maxPods: 96
kubernetesApiAccess:
- 0.0.0.0/0
kubernetesVersion: https://storage.googleapis.com/k8s-release-dev/ci/v1.35.0-alpha.3.258+183892b2c96aa9
networking:
gce: {}
nodePortAccess:
- 0.0.0.0/0
nonMasqueradeCIDR: 10.64.0.0/10
podCIDR: 10.64.0.0/10
project: k8s-infra-e2e-scale-5k-project
serviceClusterIPRange: 10.130.0.0/15
sshAccess:
- 0.0.0.0/0
subnets:
- cidr: 10.128.0.0/15
name: us-east1
region: us-east1
type: Private
topology:
dns:
type: None8. Please run the commands with most verbose logging by adding the -v 10 flag.
Paste the logs into this report, or in a gist and provide the gist link here.
9. Anything else do we need to know?