Run 37138 Add VPA support for KAI-Scheduler

CHANGELOG.md

@@ -6,6 +6,13 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/).
 
 ## [Unreleased]
 
+## [v0.15.0] - 2026-03-05
+### Added
+- Added support for VPA configuration for the different components of the KAI Scheduler - [jrosenboimnvidia](https://github.com/NVIDIA/KAI-Scheduler/pull/1119)
+- Users that have VPA installed on their cluster can now utilize it for proper vertical autoscaling
+
+
+## [v0.14.0] - 2026-03-02
 ### Fixed
 
 - Updated resource enumeration logic to exclude resources with count of 0. [#1120](https://github.com/NVIDIA/KAI-Scheduler/issues/1120)

cmd/operator/app/app.go

@@ -8,6 +8,7 @@ import (
 
 	nvidiav1 "github.com/NVIDIA/gpu-operator/api/nvidia/v1"
 	monitoringv1 "github.com/prometheus-operator/prometheus-operator/pkg/apis/monitoring/v1"
+	vpav1 "k8s.io/autoscaler/vertical-pod-autoscaler/pkg/apis/autoscaling.k8s.io/v1"
 
 	"github.com/NVIDIA/KAI-scheduler/cmd/operator/config"
 	kaiv1 "github.com/NVIDIA/KAI-scheduler/pkg/apis/kai/v1"
@@ -45,6 +46,7 @@ func init() {
 	utilruntime.Must(kaiv1alpha1.AddToScheme(scheme))
 	utilruntime.Must(nvidiav1.AddToScheme(scheme))
 	utilruntime.Must(monitoringv1.AddToScheme(scheme))
+	utilruntime.Must(vpav1.AddToScheme(scheme))
 	// +kubebuilder:scaffold:scheme
 }
 

deployments/kai-scheduler/templates/kai-config.yaml

@@ -43,6 +43,10 @@ spec:
     imagesPullSecret: {{ index .Values.global.imagePullSecrets 0 | default "" }}
     {{- end }}
     replicaCount: {{ .Values.operator.replicaCount | default 1 }}
+    {{- if .Values.global.vpa }}
+    vpa:
+      {{- toYaml .Values.global.vpa | nindent 6 }}
+    {{- end }}
 
   binder:
     service:

deployments/kai-scheduler/templates/rbac/operator.yaml

@@ -92,6 +92,18 @@ rules:
   - patch
   - update
   - watch
+- apiGroups:
+  - autoscaling.k8s.io
+  resources:
+  - verticalpodautoscalers
+  verbs:
+  - create
+  - delete
+  - get
+  - list
+  - patch
+  - update
+  - watch
 - apiGroups:
   - coordination.k8s.io
   resources:

deployments/kai-scheduler/values.yaml

@@ -16,6 +16,20 @@ global:
   tolerations: []
   namespaceLabelSelector: {}
   podLabelSelector: {}
+  vpa:
+    enabled: false
+    updatePolicy:
+      updateMode: InPlaceOrRecreate
+      minReplicas: 1
+    resourcePolicy:
+      containerPolicies:
+        - containerName: "*"
+          minAllowed:
+            cpu: 50m
+            memory: 500Mi
+          maxAllowed:
+            cpu: 2
+            memory: 5Gi
   resourceReservation:
     namespace: kai-resource-reservation
     serviceAccount: kai-resource-reservation

go.mod

@@ -46,6 +46,7 @@ require (
 	k8s.io/apiextensions-apiserver v0.34.3
 	k8s.io/apimachinery v0.34.3
 	k8s.io/apiserver v0.34.3
+	k8s.io/autoscaler/vertical-pod-autoscaler v1.5.1
 	k8s.io/cli-runtime v0.34.1
 	k8s.io/client-go v0.34.3
 	k8s.io/cluster-bootstrap v0.34.1
@@ -98,15 +99,15 @@ require (
 	github.com/cyphar/filepath-securejoin v0.6.0 // indirect
 	github.com/davecgh/go-spew v1.1.2-0.20180830191138-d8f796af33cc // indirect
 	github.com/distribution/reference v0.6.0 // indirect
-	github.com/emicklei/go-restful/v3 v3.12.2 // indirect
+	github.com/emicklei/go-restful/v3 v3.13.0 // indirect
 	github.com/evanphx/json-patch/v5 v5.9.11 // indirect
 	github.com/felixge/httpsnoop v1.0.4 // indirect
 	github.com/fsnotify/fsnotify v1.9.0 // indirect
 	github.com/fxamacker/cbor/v2 v2.9.0 // indirect
 	github.com/gabriel-vasile/mimetype v1.4.7 // indirect
 	github.com/gin-contrib/sse v0.1.0 // indirect
 	github.com/go-logr/zapr v1.3.0 // indirect
-	github.com/go-openapi/jsonpointer v0.21.1 // indirect
+	github.com/go-openapi/jsonpointer v0.21.2 // indirect
 	github.com/go-openapi/jsonreference v0.21.0 // indirect
 	github.com/go-openapi/swag v0.23.1 // indirect
 	github.com/go-playground/locales v0.14.1 // indirect
@@ -148,7 +149,7 @@ require (
 	github.com/pelletier/go-toml/v2 v2.2.3 // indirect
 	github.com/pmezard/go-difflib v1.0.1-0.20181226105442-5d4384ee4fb2 // indirect
 	github.com/prometheus/client_model v0.6.2 // indirect
-	github.com/prometheus/procfs v0.16.1 // indirect
+	github.com/prometheus/procfs v0.17.0 // indirect
 	github.com/robfig/cron/v3 v3.0.1 // indirect
 	github.com/sirupsen/logrus v1.9.3 // indirect
 	github.com/spf13/cobra v1.10.1 // indirect
@@ -177,19 +178,19 @@ require (
 	golang.org/x/sys v0.40.0 // indirect
 	golang.org/x/term v0.37.0 // indirect
 	golang.org/x/text v0.31.0 // indirect
-	golang.org/x/time v0.11.0 // indirect
+	golang.org/x/time v0.12.0 // indirect
 	golang.org/x/tools v0.38.0 // indirect
 	google.golang.org/genproto v0.0.0-20250303144028-a0af3efb3deb // indirect
 	google.golang.org/genproto/googleapis/api v0.0.0-20250303144028-a0af3efb3deb // indirect
 	google.golang.org/genproto/googleapis/rpc v0.0.0-20250313205543-e70fdf4c4cb4 // indirect
 	google.golang.org/protobuf v1.36.8 // indirect
-	gopkg.in/evanphx/json-patch.v4 v4.12.0 // indirect
+	gopkg.in/evanphx/json-patch.v4 v4.13.0 // indirect
 	gopkg.in/inf.v0 v0.9.1 // indirect
 	k8s.io/cloud-provider v0.34.1 // indirect
 	k8s.io/controller-manager v0.34.1 // indirect
 	k8s.io/cri-api v0.34.1 // indirect
 	k8s.io/csi-translation-lib v0.34.1 // indirect
-	k8s.io/kube-openapi v0.0.0-20250710124328-f3f2b991d03b // indirect
+	k8s.io/kube-openapi v0.0.0-20250814151709-d7b6acb124c3 // indirect
 	k8s.io/kubelet v0.34.1 // indirect
 	knative.dev/networking v0.0.0-20250117155906-67d1c274ba6a // indirect
 	sigs.k8s.io/json v0.0.0-20250730193827-2d320260d730 // indirect

go.sum

@@ -71,8 +71,8 @@ github.com/distribution/reference v0.6.0 h1:0IXCQ5g4/QMHHkarYzh5l+u8T3t73zM5Qvfr
 github.com/distribution/reference v0.6.0/go.mod h1:BbU0aIcezP1/5jX/8MP0YiH4SdvB5Y4f/wlDRiLyi3E=
 github.com/dustin/go-humanize v1.0.1 h1:GzkhY7T5VNhEkwH0PVJgjz+fX1rhBrR7pRT3mDkpeCY=
 github.com/dustin/go-humanize v1.0.1/go.mod h1:Mu1zIs6XwVuF/gI1OepvI0qD18qycQx+mFykh5fBlto=
-github.com/emicklei/go-restful/v3 v3.12.2 h1:DhwDP0vY3k8ZzE0RunuJy8GhNpPL6zqLkDf9B/a0/xU=
-github.com/emicklei/go-restful/v3 v3.12.2/go.mod h1:6n3XBCmQQb25CM2LCACGz8ukIrRry+4bhvbpWn3mrbc=
+github.com/emicklei/go-restful/v3 v3.13.0 h1:C4Bl2xDndpU6nJ4bc1jXd+uTmYPVUwkD6bFY/oTyCes=
+github.com/emicklei/go-restful/v3 v3.13.0/go.mod h1:6n3XBCmQQb25CM2LCACGz8ukIrRry+4bhvbpWn3mrbc=
 github.com/envoyproxy/go-control-plane v0.9.0/go.mod h1:YTl/9mNaCwkRvm6d1a2C3ymFceY/DCBVvsKhRF0iEA4=
 github.com/envoyproxy/go-control-plane v0.9.1-0.20191026205805-5f8ba28d4473/go.mod h1:YTl/9mNaCwkRvm6d1a2C3ymFceY/DCBVvsKhRF0iEA4=
 github.com/envoyproxy/go-control-plane v0.9.4/go.mod h1:6rpuAdCZL397s3pYoYcLgu1mIlRU8Am5FuJP05cCM98=
@@ -107,8 +107,8 @@ github.com/go-logr/stdr v1.2.2 h1:hSWxHoqTgW2S2qGc0LTAI563KZ5YKYRhT3MFKZMbjag=
 github.com/go-logr/stdr v1.2.2/go.mod h1:mMo/vtBO5dYbehREoey6XUKy/eSumjCCveDpRre4VKE=
 github.com/go-logr/zapr v1.3.0 h1:XGdV8XW8zdwFiwOA2Dryh1gj2KRQyOOoNmBy4EplIcQ=
 github.com/go-logr/zapr v1.3.0/go.mod h1:YKepepNBd1u/oyhd/yQmtjVXmm9uML4IXUgMOwR8/Gg=
-github.com/go-openapi/jsonpointer v0.21.1 h1:whnzv/pNXtK2FbX/W9yJfRmE2gsmkfahjMKB0fZvcic=
-github.com/go-openapi/jsonpointer v0.21.1/go.mod h1:50I1STOfbY1ycR8jGz8DaMeLCdXiI6aDteEdRNNzpdk=
+github.com/go-openapi/jsonpointer v0.21.2 h1:AqQaNADVwq/VnkCmQg6ogE+M3FOsKTytwges0JdwVuA=
+github.com/go-openapi/jsonpointer v0.21.2/go.mod h1:50I1STOfbY1ycR8jGz8DaMeLCdXiI6aDteEdRNNzpdk=
 github.com/go-openapi/jsonreference v0.21.0 h1:Rs+Y7hSXT83Jacb7kFyjn4ijOuVGSvOdF2+tg1TRrwQ=
 github.com/go-openapi/jsonreference v0.21.0/go.mod h1:LmZmgsrTkVg9LG4EaHeY8cBDslNPMo06cago5JNLkm4=
 github.com/go-openapi/swag v0.23.1 h1:lpsStH0n2ittzTnbaSloVZLuB5+fvSY/+hnagBjSNZU=
@@ -257,8 +257,8 @@ github.com/prometheus/client_model v0.6.2 h1:oBsgwpGs7iVziMvrGhE53c/GrLUsZdHnqNw
 github.com/prometheus/client_model v0.6.2/go.mod h1:y3m2F6Gdpfy6Ut/GBsUqTWZqCUvMVzSfMLjcu6wAwpE=
 github.com/prometheus/common v0.66.1 h1:h5E0h5/Y8niHc5DlaLlWLArTQI7tMrsfQjHV+d9ZoGs=
 github.com/prometheus/common v0.66.1/go.mod h1:gcaUsgf3KfRSwHY4dIMXLPV0K/Wg1oZ8+SbZk/HH/dA=
-github.com/prometheus/procfs v0.16.1 h1:hZ15bTNuirocR6u0JZ6BAHHmwS1p8B4P6MRqxtzMyRg=
-github.com/prometheus/procfs v0.16.1/go.mod h1:teAbpZRB1iIAJYREa1LsoWUXykVXA1KlTmWl8x/U+Is=
+github.com/prometheus/procfs v0.17.0 h1:FuLQ+05u4ZI+SS/w9+BWEM2TXiHKsUQ9TADiRH7DuK0=
+github.com/prometheus/procfs v0.17.0/go.mod h1:oPQLaDAMRbA+u8H5Pbfq+dl3VDAvHxMUOVhe0wYB2zw=
 github.com/prometheus/statsd_exporter v0.22.7 h1:7Pji/i2GuhK6Lu7DHrtTkFmNBCudCPT1pX2CziuyQR0=
 github.com/prometheus/statsd_exporter v0.22.7/go.mod h1:N/TevpjkIh9ccs6nuzY3jQn9dFqnUakOjnEuMPJJJnI=
 github.com/ray-project/kuberay/ray-operator v1.4.2 h1:A4tGzbIky8sInAUxZBdBb+rrpZ7fbqoxdsOtm559Zqg=
@@ -403,8 +403,8 @@ golang.org/x/text v0.3.0/go.mod h1:NqM8EUOU14njkJ3fqMW+pc6Ldnwhi/IjpwHt7yyuwOQ=
 golang.org/x/text v0.3.3/go.mod h1:5Zoc/QRtKVWzQhOtBMvqHzDpF6irO9z98xDceosuGiQ=
 golang.org/x/text v0.31.0 h1:aC8ghyu4JhP8VojJ2lEHBnochRno1sgL6nEi9WGFGMM=
 golang.org/x/text v0.31.0/go.mod h1:tKRAlv61yKIjGGHX/4tP1LTbc13YSec1pxVEWXzfoeM=
-golang.org/x/time v0.11.0 h1:/bpjEDfN9tkoN/ryeYHnv5hcMlc8ncjMcM4XBk5NWV0=
-golang.org/x/time v0.11.0/go.mod h1:CDIdPxbZBQxdj6cxyCIdrNogrJKMJ7pr37NYpMcMDSg=
+golang.org/x/time v0.12.0 h1:ScB/8o8olJvc+CQPWrK3fPZNfh7qgwCrY0zJmoEQLSE=
+golang.org/x/time v0.12.0/go.mod h1:CDIdPxbZBQxdj6cxyCIdrNogrJKMJ7pr37NYpMcMDSg=
 golang.org/x/tools v0.0.0-20180917221912-90fa682c2a6e/go.mod h1:n7NCudcB/nEzxVGmLbDWY5pfWTLqBcC2KZ6jyYvM4mQ=
 golang.org/x/tools v0.0.0-20190114222345-bf090417da8b/go.mod h1:n7NCudcB/nEzxVGmLbDWY5pfWTLqBcC2KZ6jyYvM4mQ=
 golang.org/x/tools v0.0.0-20190226205152-f727befe758c/go.mod h1:9Yl7xja0Znq3iFh3HoIrodX9oNMXvdceNzlUR8zjMvY=
@@ -446,8 +446,8 @@ google.golang.org/protobuf v1.36.8/go.mod h1:fuxRtAxBytpl4zzqUh6/eyUujkJdNiuEkXn
 gopkg.in/check.v1 v0.0.0-20161208181325-20d25e280405/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=
 gopkg.in/check.v1 v1.0.0-20201130134442-10cb98267c6c h1:Hei/4ADfdWqJk1ZMxUNpqntNwaWcugrBjAiHlqqRiVk=
 gopkg.in/check.v1 v1.0.0-20201130134442-10cb98267c6c/go.mod h1:JHkPIbrfpd72SG/EVd6muEfDQjcINNoR0C8j2r3qZ4Q=
-gopkg.in/evanphx/json-patch.v4 v4.12.0 h1:n6jtcsulIzXPJaxegRbvFNNrZDjbij7ny3gmSPG+6V4=
-gopkg.in/evanphx/json-patch.v4 v4.12.0/go.mod h1:p8EYWUEYMpynmqDbY58zCKCFZw8pRWMG4EsWvDvM72M=
+gopkg.in/evanphx/json-patch.v4 v4.13.0 h1:czT3CmqEaQ1aanPc5SdlgQrrEIb8w/wwCvWWnfEbYzo=
+gopkg.in/evanphx/json-patch.v4 v4.13.0/go.mod h1:p8EYWUEYMpynmqDbY58zCKCFZw8pRWMG4EsWvDvM72M=
 gopkg.in/h2non/gock.v1 v1.1.2 h1:jBbHXgGBK/AoPVfJh5x4r/WxIrElvbLel8TCZkkZJoY=
 gopkg.in/h2non/gock.v1 v1.1.2/go.mod h1:n7UGz/ckNChHiK05rDoiC4MYSunEC/lyaUm2WWaDva0=
 gopkg.in/inf.v0 v0.9.1 h1:73M5CoZyi3ZLMOyDlQh031Cx6N9NDJ2Vvfl76EDAgDc=
@@ -470,6 +470,8 @@ k8s.io/apimachinery v0.34.3 h1:/TB+SFEiQvN9HPldtlWOTp0hWbJ+fjU+wkxysf/aQnE=
 k8s.io/apimachinery v0.34.3/go.mod h1:/GwIlEcWuTX9zKIg2mbw0LRFIsXwrfoVxn+ef0X13lw=
 k8s.io/apiserver v0.34.3 h1:uGH1qpDvSiYG4HVFqc6A3L4CKiX+aBWDrrsxHYK0Bdo=
 k8s.io/apiserver v0.34.3/go.mod h1:QPnnahMO5C2m3lm6fPW3+JmyQbvHZQ8uudAu/493P2w=
+k8s.io/autoscaler/vertical-pod-autoscaler v1.5.1 h1:LlVtM3IKqIVHz1ZXC3ahe/mAtDWb7Eob0tyTzqFULqg=
+k8s.io/autoscaler/vertical-pod-autoscaler v1.5.1/go.mod h1:znhUnV0Yn+CkZu3TZ2HVqd8GFRMkPj/CXszX1gdBjTU=
 k8s.io/cli-runtime v0.34.1 h1:btlgAgTrYd4sk8vJTRG6zVtqBKt9ZMDeQZo2PIzbL7M=
 k8s.io/cli-runtime v0.34.1/go.mod h1:aVA65c+f0MZiMUPbseU/M9l1Wo2byeaGwUuQEQVVveE=
 k8s.io/client-go v0.34.3 h1:wtYtpzy/OPNYf7WyNBTj3iUA0XaBHVqhv4Iv3tbrF5A=
@@ -502,8 +504,8 @@ k8s.io/kube-aggregator v0.34.1 h1:WNLV0dVNoFKmuyvdWLd92iDSyD/TSTjqwaPj0U9XAEU=
 k8s.io/kube-aggregator v0.34.1/go.mod h1:RU8j+5ERfp0h+gIvWtxRPfsa5nK7rboDm8RST8BJfYQ=
 k8s.io/kube-controller-manager v0.34.1 h1:hrPRR4toT+xABAxzGpnldTL1RocYXyVhx6A5Einb9wU=
 k8s.io/kube-controller-manager v0.34.1/go.mod h1:+7jKjj5i7NLGM6zPHbdMh7qHaWFOBsF/oeUDdS70DSg=
-k8s.io/kube-openapi v0.0.0-20250710124328-f3f2b991d03b h1:MloQ9/bdJyIu9lb1PzujOPolHyvO06MXG5TUIj2mNAA=
-k8s.io/kube-openapi v0.0.0-20250710124328-f3f2b991d03b/go.mod h1:UZ2yyWbFTpuhSbFhv24aGNOdoRdJZgsIObGBUaYVsts=
+k8s.io/kube-openapi v0.0.0-20250814151709-d7b6acb124c3 h1:liMHz39T5dJO1aOKHLvwaCjDbf07wVh6yaUlTpunnkE=
+k8s.io/kube-openapi v0.0.0-20250814151709-d7b6acb124c3/go.mod h1:UZ2yyWbFTpuhSbFhv24aGNOdoRdJZgsIObGBUaYVsts=
 k8s.io/kube-proxy v0.34.1 h1:cIriNCJY5XmRhXCCyQiazyqi47lbwcBQf0H76fVOpkw=
 k8s.io/kube-proxy v0.34.1/go.mod h1:syed9c5+gUVFMo6p24SnlTHzsp+BMd4ACcTw2dbArw0=
 k8s.io/kube-scheduler v0.34.1 h1:S5td6VZwC3lCqERXclerDXhJ26zYc6JroY0s03+PqJ8=

hack/setup-e2e-cluster.sh

@@ -18,6 +18,7 @@ KIND_CONFIG=${REPO_ROOT}/hack/e2e-kind-config.yaml
 # Parse named parameters
 TEST_THIRD_PARTY_INTEGRATIONS=${TEST_THIRD_PARTY_INTEGRATIONS:-"false"}
 LOCAL_IMAGES_BUILD=${LOCAL_IMAGES_BUILD:-"false"}
+INSTALL_VPA=${INSTALL_VPA:-"false"}
 
 while [[ $# -gt 0 ]]; do
   case $1 in
@@ -29,10 +30,15 @@ while [[ $# -gt 0 ]]; do
       LOCAL_IMAGES_BUILD="true"
       shift
       ;;
+    --install-vpa)
+      INSTALL_VPA="true"
+      shift
+      ;;
     -h|--help)
-      echo "Usage: $0 [--test-third-party-integrations] [--local-images-build]"
+      echo "Usage: $0 [--test-third-party-integrations] [--local-images-build] [--install-vpa]"
       echo "  --test-third-party-integrations: Install third party operators for compatibility testing"
       echo "  --local-images-build: Build and use local images instead of pulling from registry"
+      echo "  --install-vpa: Install Vertical Pod Autoscaler and metrics-server"
       exit 0
       ;;
     *)
@@ -67,6 +73,23 @@ helm install prometheus prometheus-community/kube-prometheus-stack --namespace m
     --set "prometheus.enabled=false" \
     --wait
 
+# Install VPA and its prerequisites
+if [ "$INSTALL_VPA" = "true" ]; then
+    echo "Installing metrics-server (required by VPA recommender)..."
+    kubectl apply -f https://github.com/kubernetes-sigs/metrics-server/releases/download/v0.8.1/components.yaml
+    # kind uses self-signed kubelet certs, so metrics-server needs --kubelet-insecure-tls
+    kubectl patch deployment metrics-server -n kube-system --type=json \
+        -p '[{"op":"add","path":"/spec/template/spec/containers/0/args/-","value":"--kubelet-insecure-tls"}]'
+    kubectl wait --for=condition=available --timeout=120s deployment/metrics-server -n kube-system
+
+    echo "Installing Vertical Pod Autoscaler..."
+    VPA_TMPDIR=$(mktemp -d)
+    git clone https://github.com/kubernetes/autoscaler.git "$VPA_TMPDIR/autoscaler"
+    (cd "$VPA_TMPDIR/autoscaler/vertical-pod-autoscaler" && git checkout vertical-pod-autoscaler-1.5.1 && ./hack/vpa-up.sh)
+    rm -rf "$VPA_TMPDIR"
+    echo "VPA installation complete."
+fi
+
 # Install third party operators to check the compatibility with the kai-scheduler
 if [ "$TEST_THIRD_PARTY_INTEGRATIONS" = "true" ]; then
     ${REPO_ROOT}/hack/third_party_integrations/deploy_ray.sh

pkg/apis/kai/v1/admission/admission.go

@@ -48,9 +48,13 @@ type Admission struct {
 	// set to empty string to disable
 	// +kubebuilder:validation:Optional
 	GPUPodRuntimeClassName *string `json:"gpuPodRuntimeClassName,omitempty"`
+
+	// VPA specifies Vertical Pod Autoscaler configuration for the admission service
+	// +kubebuilder:validation:Optional
+	VPA *common.VPASpec `json:"vpa,omitempty"`
 }
 
-func (b *Admission) SetDefaultsWhereNeeded(replicaCount *int32) {
+func (b *Admission) SetDefaultsWhereNeeded(replicaCount *int32, globalVPA *common.VPASpec) {
 	b.Service = common.SetDefault(b.Service, &common.Service{})
 	b.Service.SetDefaultsWhereNeeded(imageName)
 
@@ -68,6 +72,10 @@ func (b *Admission) SetDefaultsWhereNeeded(replicaCount *int32) {
 	b.MutatingWebhookConfigurationName = common.SetDefault(b.MutatingWebhookConfigurationName, ptr.To(defaultMutatingWebhookName))
 
 	b.GPUPodRuntimeClassName = common.SetDefault(b.GPUPodRuntimeClassName, ptr.To(constants.DefaultRuntimeClassName))
+
+	if b.VPA == nil {
+		b.VPA = globalVPA
+	}
 }
 
 // Webhook defines configuration for the admission webhook

pkg/apis/kai/v1/admission/admission_test.go

@@ -22,7 +22,7 @@ var _ = Describe("Admission", func() {
 		Admission := &Admission{}
 		var replicaCount int32
 		replicaCount = 1
-		Admission.SetDefaultsWhereNeeded(&replicaCount)
+		Admission.SetDefaultsWhereNeeded(&replicaCount, nil)
 		Expect(*Admission.Service.Enabled).To(Equal(true))
 		Expect(*Admission.Service.Image.Name).To(Equal("admission"))
 		Expect(*Admission.Replicas).To(Equal(int32(1)))
@@ -32,7 +32,7 @@ var _ = Describe("Admission", func() {
 		Admission := &Admission{}
 		var replicaCount int32
 		replicaCount = 3
-		Admission.SetDefaultsWhereNeeded(&replicaCount)
+		Admission.SetDefaultsWhereNeeded(&replicaCount, nil)
 		Expect(*Admission.Replicas).To(Equal(int32(3)))
 	})
 })

pkg/apis/kai/v1/binder/binder.go

@@ -47,9 +47,13 @@ type Binder struct {
 	// leave empty if unsure to let the operator auto detect using ClusterPolicy (nvidia gpu-operator only)
 	// +kubebuilder:validation:Optional
 	CDIEnabled *bool `json:"cdiEnabled,omitempty"`
+
+	// VPA specifies Vertical Pod Autoscaler configuration for the binder
+	// +kubebuilder:validation:Optional
+	VPA *common.VPASpec `json:"vpa,omitempty"`
 }
 
-func (b *Binder) SetDefaultsWhereNeeded(replicaCount *int32) {
+func (b *Binder) SetDefaultsWhereNeeded(replicaCount *int32, globalVPA *common.VPASpec) {
 	b.Service = common.SetDefault(b.Service, &common.Service{})
 	b.Service.Resources = common.SetDefault(b.Service.Resources, &common.Resources{})
 	if b.Service.Resources.Requests == nil {
@@ -81,6 +85,10 @@ func (b *Binder) SetDefaultsWhereNeeded(replicaCount *int32) {
 
 	b.ProbePort = common.SetDefault(b.ProbePort, ptr.To(8081))
 	b.MetricsPort = common.SetDefault(b.MetricsPort, ptr.To(8080))
+
+	if b.VPA == nil {
+		b.VPA = globalVPA
+	}
 }
 
 type ResourceReservation struct {

pkg/apis/kai/v1/binder/binder_test.go

@@ -23,7 +23,7 @@ func TestBinder(t *testing.T) {
 var _ = Describe("Binder", func() {
 	It("Set Defaults", func(ctx context.Context) {
 		binder := &Binder{}
-		binder.SetDefaultsWhereNeeded(nil)
+		binder.SetDefaultsWhereNeeded(nil, nil)
 		Expect(*binder.Service.Enabled).To(Equal(true))
 		Expect(*binder.Service.Image.Name).To(Equal("binder"))
 		Expect(binder.Service.Resources.Requests[v1.ResourceCPU]).To(Equal(resource.MustParse("50m")))
@@ -35,14 +35,14 @@ var _ = Describe("Binder", func() {
 		binder := &Binder{}
 		var replicaCount int32
 		replicaCount = 3
-		binder.SetDefaultsWhereNeeded(&replicaCount)
+		binder.SetDefaultsWhereNeeded(&replicaCount, nil)
 		Expect(*binder.Replicas).To(Equal(int32(3)))
 	})
 
 	Context("ResourceReservation PodResources configuration", func() {
 		It("should not set default PodResources when not configured", func(ctx context.Context) {
 			binder := &Binder{}
-			binder.SetDefaultsWhereNeeded(nil)
+			binder.SetDefaultsWhereNeeded(nil, nil)
 
 			// PodResources should be nil when not configured
 			Expect(binder.ResourceReservation.PodResources).To(BeNil())
@@ -64,7 +64,7 @@ var _ = Describe("Binder", func() {
 					PodResources: podResources,
 				},
 			}
-			binder.SetDefaultsWhereNeeded(nil)
+			binder.SetDefaultsWhereNeeded(nil, nil)
 
 			// Configured values should be preserved
 			Expect(binder.ResourceReservation.PodResources).NotTo(BeNil())
@@ -88,7 +88,7 @@ var _ = Describe("Binder", func() {
 					PodResources: podResources,
 				},
 			}
-			binder.SetDefaultsWhereNeeded(nil)
+			binder.SetDefaultsWhereNeeded(nil, nil)
 
 			// Only CPU should be set
 			Expect(binder.ResourceReservation.PodResources).NotTo(BeNil())