Merge pull request llm-d#451 from llm-d-incubation/runner-ci-policy

Add ok-to-test gate, use cluster HF token for e2e tests, fix istio issues, run wva against 2 different stacks simultaneously

Author: clubanderson

.github/workflows/ci-e2e-openshift.yaml

@@ -36,12 +36,10 @@ jobs:
 
       # - name: Run precommit checks
       #   run: make precommit
-      
-      - name: Run make test-e2e
-        shell: bash
-        run: |
-          make test-e2e
-      
+
+      # Note: test-e2e requires Kind and full cluster setup
+      # E2E tests run in separate ci-e2e-openshift workflow on self-hosted runner
+
       - name: Run make build
         shell: bash
         run: |

.github/workflows/ci-pr-checks.yaml

@@ -142,6 +142,36 @@ spec:
 
 More examples in [config/samples/](config/samples/).
 
+## Upgrading
+
+### CRD Updates
+
+**Important:** Helm does not automatically update CRDs during `helm upgrade`. When upgrading WVA to a new version with CRD changes, you must manually apply the updated CRDs first:
+
+```bash
+# Apply the latest CRDs before upgrading
+kubectl apply -f charts/workload-variant-autoscaler/crds/
+
+# Then upgrade the Helm release
+helm upgrade workload-variant-autoscaler ./charts/workload-variant-autoscaler \
+  --namespace workload-variant-autoscaler-system \
+  [your-values...]
+```
+
+### Breaking Changes
+
+#### v0.5.0 (upcoming)
+- **VariantAutoscaling CRD**: Added `scaleTargetRef` field to explicitly specify the target deployment. If not set, the controller infers the target from the `modelID` field.
+
+### Verifying CRD Version
+
+To check if your cluster has the latest CRD schema:
+
+```bash
+# Check the CRD fields
+kubectl get crd variantautoscalings.llmd.ai -o jsonpath='{.spec.versions[0].schema.openAPIV3Schema.properties.spec.properties}' | jq 'keys'
+```
+
 ## Contributing
 
 We welcome contributions! See the llm-d Contributing Guide for guidelines.

README.md

@@ -0,0 +1,58 @@
+{{/*
+Expand the name of the chart.
+*/}}
+{{- define "workload-variant-autoscaler.name" -}}
+{{- default .Chart.Name .Values.nameOverride | trunc 63 | trimSuffix "-" }}
+{{- end }}
+
+{{/*
+Create a default fully qualified app name.
+We truncate at 63 chars because some Kubernetes name fields are limited to this (by the DNS naming spec).
+If release name contains chart name it will be used as a full name.
+*/}}
+{{- define "workload-variant-autoscaler.fullname" -}}
+{{- if .Values.fullnameOverride }}
+{{- .Values.fullnameOverride | trunc 63 | trimSuffix "-" }}
+{{- else }}
+{{- $name := default .Chart.Name .Values.nameOverride }}
+{{- if contains $name .Release.Name }}
+{{- .Release.Name | trunc 63 | trimSuffix "-" }}
+{{- else }}
+{{- printf "%s-%s" .Release.Name $name | trunc 63 | trimSuffix "-" }}
+{{- end }}
+{{- end }}
+{{- end }}
+
+{{/*
+Create chart name and version as used by the chart label.
+*/}}
+{{- define "workload-variant-autoscaler.chart" -}}
+{{- printf "%s-%s" .Chart.Name .Chart.Version | replace "+" "_" | trunc 63 | trimSuffix "-" }}
+{{- end }}
+
+{{/*
+Common labels
+*/}}
+{{- define "workload-variant-autoscaler.labels" -}}
+helm.sh/chart: {{ include "workload-variant-autoscaler.chart" . }}
+{{ include "workload-variant-autoscaler.selectorLabels" . }}
+{{- if .Chart.AppVersion }}
+app.kubernetes.io/version: {{ .Chart.AppVersion | quote }}
+{{- end }}
+app.kubernetes.io/managed-by: {{ .Release.Service }}
+{{- end }}
+
+{{/*
+Selector labels
+*/}}
+{{- define "workload-variant-autoscaler.selectorLabels" -}}
+app.kubernetes.io/name: {{ include "workload-variant-autoscaler.name" . }}
+app.kubernetes.io/instance: {{ .Release.Name }}
+{{- end }}
+
+{{/*
+Create the name of the service account to use
+*/}}
+{{- define "workload-variant-autoscaler.serviceAccountName" -}}
+{{- default (include "workload-variant-autoscaler.fullname" .) .Values.serviceAccount.name }}
+{{- end }}

charts/workload-variant-autoscaler/templates/_helpers.tpl

@@ -2,8 +2,10 @@
 apiVersion: autoscaling/v2
 kind: HorizontalPodAutoscaler
 metadata:
-  name: vllm-deployment-hpa
+  name: {{ include "workload-variant-autoscaler.fullname" . }}-hpa
   namespace: {{ .Values.llmd.namespace }}
+  labels:
+    {{- include "workload-variant-autoscaler.labels" . | nindent 4 }}
 spec:
   scaleTargetRef:
     apiVersion: apps/v1
@@ -34,6 +36,7 @@ spec:
         selector:
           matchLabels:
             variant_name: {{ printf "%s-decode" .Values.llmd.modelName }}
+            exported_namespace: {{ .Values.llmd.namespace }}
       target:
         type: AverageValue
         averageValue: {{ .Values.hpa.targetAverageValue | quote }}

charts/workload-variant-autoscaler/templates/hpa.yaml

@@ -1,10 +1,10 @@
+{{- if .Values.controller.enabled }}
 apiVersion: rbac.authorization.k8s.io/v1
 kind: ClusterRoleBinding
 metadata:
-  name: prometheus-adapter-monitoring
+  name: {{ include "workload-variant-autoscaler.fullname" . }}-prometheus-adapter-monitoring
   labels:
-    app.kubernetes.io/name: workload-variant-autoscaler
-    app.kubernetes.io/instance: {{ .Release.Name }}
+    {{- include "workload-variant-autoscaler.labels" . | nindent 4 }}
 roleRef:
   apiGroup: rbac.authorization.k8s.io
   kind: ClusterRole
@@ -13,3 +13,4 @@ subjects:
   - kind: ServiceAccount
     name: {{ .Values.wva.prometheus.serviceAccountName }}
     namespace: {{ .Values.wva.prometheus.monitoringNamespace }}
+{{- end }}

charts/workload-variant-autoscaler/templates/manager/prometheus-clusterrolebinding.yaml

@@ -1,15 +1,16 @@
+{{- if .Values.controller.enabled }}
 apiVersion: rbac.authorization.k8s.io/v1
 kind: ClusterRoleBinding
 metadata:
-  name: workload-variant-autoscaler-monitoring
+  name: {{ include "workload-variant-autoscaler.fullname" . }}-monitoring
   labels:
-    app.kubernetes.io/name: workload-variant-autoscaler
-    app.kubernetes.io/instance: {{ .Release.Name }}
+    {{- include "workload-variant-autoscaler.labels" . | nindent 4 }}
 roleRef:
   apiGroup: rbac.authorization.k8s.io
   kind: ClusterRole
   name: cluster-monitoring-view
 subjects:
   - kind: ServiceAccount
-    name: workload-variant-autoscaler-controller-manager
+    name: {{ include "workload-variant-autoscaler.fullname" . }}-controller-manager
     namespace: {{ .Release.Namespace }}
+{{- end }}

charts/workload-variant-autoscaler/templates/manager/wva-clusterrolebinding.yaml

@@ -1,8 +1,11 @@
+{{- if .Values.controller.enabled }}
 apiVersion: v1
 kind: ConfigMap
 metadata:
-  name: accelerator-unit-costs
+  name: {{ include "workload-variant-autoscaler.fullname" . }}-accelerator-unit-costs
   namespace: {{ .Release.Namespace }}
+  labels:
+    {{- include "workload-variant-autoscaler.labels" . | nindent 4 }}
 data:
   A100: |
     {
@@ -34,3 +37,4 @@ data:
     "device": "NVIDIA-L40S",
     "cost": "32.00"
     }
+{{- end }}

charts/workload-variant-autoscaler/templates/manager/wva-configmap-accelerator-costs.yaml

@@ -1,3 +1,4 @@
+{{- if .Values.controller.enabled }}
 apiVersion: v1
 kind: ConfigMap
 # This configMap defines the set of accelerators available
@@ -9,8 +10,10 @@ kind: ConfigMap
 # - cost is the cents/hour cost of this accelerator
 #
 metadata:
-  name: service-classes-config
+  name: {{ include "workload-variant-autoscaler.fullname" . }}-service-classes-config
   namespace: {{ .Release.Namespace }}
+  labels:
+    {{- include "workload-variant-autoscaler.labels" . | nindent 4 }}
 data:
   premium.yaml: |
     name: Premium
@@ -35,3 +38,4 @@ data:
       - model: meta/llama0-7b
         slo-tpot: 150
         slo-ttft: 1500
+{{- end }}

charts/workload-variant-autoscaler/templates/manager/wva-configmap-service-class.yaml

@@ -1,10 +1,11 @@
+{{- if .Values.controller.enabled }}
 apiVersion: v1
 kind: ConfigMap
 metadata:
-  name: workload-variant-autoscaler-variantautoscaling-config
+  name: {{ include "workload-variant-autoscaler.fullname" . }}-variantautoscaling-config
   namespace: {{ .Release.Namespace }}
   labels:
-    app.kubernetes.io/name: workload-variant-autoscaler
+    {{- include "workload-variant-autoscaler.labels" . | nindent 4 }}
 data:
   # Prometheus configuration - REQUIRED: Set your Prometheus server URL
   # Examples:
@@ -21,11 +22,11 @@ data:
   # PROMETHEUS_CLIENT_KEY_PATH: "/path/to/client.key"   # Client private key for mutual TLS
   # PROMETHEUS_SERVER_NAME: "prometheus.example.com"    # Expected server name for SNI
   PROMETHEUS_TLS_INSECURE_SKIP_VERIFY: {{ if and .Values.wva.prometheus.tls (hasKey .Values.wva.prometheus.tls "insecureSkipVerify") }}{{ .Values.wva.prometheus.tls.insecureSkipVerify | quote }}{{ else }}"true"{{ end }}
-  
+
   # Authentication Configuration (BearerToken takes precedence over TokenPath)
   # PROMETHEUS_BEARER_TOKEN: "your-token-here"           # Direct bearer token (development/testing)
   # PROMETHEUS_TOKEN_PATH: "/path/to/token/file"        # Path to bearer token file (production with mounted secrets)
-  
+
   # Optimization configuration
   GLOBAL_OPT_INTERVAL: {{ .Values.wva.reconcileInterval | quote }}
 
@@ -56,3 +57,4 @@ data:
   # EPP_METRICS_CACHE_TTL: "15s"
   # EPP_METRICS_CACHE_MAX_SIZE: "500"
   # EPP_METRICS_CACHE_CLEANUP_INTERVAL: "30s"
+{{- end }}