helm support for sidecar injection in EPP (#1821)

* draft

* draft

* workable version

* added docs

* fixed format

* move epp-envoy to a standalone file

* make standalone a provider

* removed unused change

* remove unused change

* syntax fix

* workable version

* remove unused yaml

* added more customized info and removed example

* fixed typo

* fixed default image pull policy

Author: capri-xiyue

config/charts/inferencepool/templates/epp-config.yaml

@@ -24,3 +24,13 @@ data:
   {{- .Values.inferenceExtension.pluginsCustomConfig | toYaml | nindent 2 }}
   {{- end }}
 
+---
+{{- if .Values.inferenceExtension.sidecar.enabled }}
+apiVersion: v1
+kind: ConfigMap
+metadata:
+  name: {{ .Values.inferenceExtension.sidecar.configMap.name }}
+  namespace: {{ .Release.Namespace }}
+data:
+  {{- .Values.inferenceExtension.sidecar.configMap.data | toYaml | nindent 2 }}
+{{- end }}

config/charts/inferencepool/templates/epp-deployment.yaml

@@ -27,9 +27,48 @@ spec:
       # Conservatively, this timeout should mirror the longest grace period of the pods within the pool
       terminationGracePeriodSeconds: 130
       containers:
+      {{- if .Values.inferenceExtension.sidecar.enabled }}
+      - name: {{ .Values.inferenceExtension.sidecar.name }}
+        image: {{ .Values.inferenceExtension.sidecar.image }}
+        imagePullPolicy: {{ .Values.inferenceExtension.sidecar.image.pullPolicy | default "IfNotPresent" }}
+        {{- with .Values.inferenceExtension.sidecar.command }}
+        command:
+          - {{ . | quote }}
+        {{- end }}
+        {{- with .Values.inferenceExtension.sidecar.args }}
+        args:
+          {{- range . }}
+            - {{ tpl . $ | quote }}
+          {{- end }}
+        {{- end }}
+        {{- with .Values.inferenceExtension.sidecar.env }}
+        env:
+          {{- toYaml . | nindent 10 }}
+          {{- end }}
+        {{- with .Values.inferenceExtension.sidecar.ports }}
+        ports:
+          {{- toYaml . | nindent 10 }}
+        {{- end }}
+        {{- with .Values.inferenceExtension.sidecar.livenessProbe }}
+        livenessProbe:
+          {{- toYaml . | nindent 10 }}
+        {{- end }}
+        {{- with .Values.inferenceExtension.sidecar.readinessProbe }}
+        readinessProbe:
+          {{- toYaml . | nindent 10 }}
+        {{- end }}
+        {{- with .Values.inferenceExtension.sidecar.resources }}
+        resources:
+          {{- toYaml . | nindent 10 }}
+        {{- end }}
+        {{- with .Values.inferenceExtension.sidecar.volumeMounts }}
+        volumeMounts:
+          {{- toYaml . | nindent 10 }}
+        {{- end }}
+        {{- end }}
       - name: epp
         image: {{ .Values.inferenceExtension.image.hub }}/{{ .Values.inferenceExtension.image.name }}:{{ .Values.inferenceExtension.image.tag }}
-        imagePullPolicy: {{ .Values.inferenceExtension.image.pullPolicy | default "Always" }}
+        imagePullPolicy: {{ .Values.inferenceExtension.image.pullPolicy | default "IfNotPresent" }}
         args:
         - --pool-name
         - {{ .Release.Name }}
@@ -139,6 +178,9 @@ spec:
         - name: plugins-config-volume
           mountPath: "/config"
       volumes:
+      {{- if .Values.inferenceExtension.sidecar.volumes }}
+      {{- tpl (toYaml .Values.inferenceExtension.sidecar.volumes) $ | nindent 6 }}
+      {{- end }}
       - name: plugins-config-volume
         configMap:
           name: {{ include "gateway-api-inference-extension.name" . }}

config/charts/inferencepool/values.yaml

@@ -12,6 +12,11 @@ inferenceExtension:
   extraContainerPorts: []
   # Define additional service ports
   extraServicePorts: []
+#  extraServicePorts:
+#    - name: http
+#      port: 8081
+#      protocol: TCP
+#      targetPort: 8081
 
   # This is the plugins configuration file.
   # pluginsCustomConfig:
@@ -38,6 +43,10 @@ inferenceExtension:
   affinity: {}
 
   tolerations: []
+  
+  # Sidecar configuration for EPP
+  sidecar:
+    enabled: false
 
   # Monitoring configuration for EPP
   monitoring: