feat: add parameters in onprem k8s chart (volume, resources & env. variables)

deploy/k8s-onprem/templates/deployment.yaml

@@ -52,24 +52,28 @@ spec:
 
     spec:
       serviceAccountName: {{ template "triton-inference-server.fullname" . }}
+      {{- with .Values.volumes }}
       volumes:
-        - name: models
-          nfs:
-            server: {{ .Values.image.modelRepositoryServer }}
-            path: {{ .Values.image.modelRepositoryPath }}
-            readOnly: false
+      {{- toYaml . | nindent 8 }}
+      {{- end }}
       containers:
         - name: {{ .Chart.Name }}
           image: "{{ .Values.image.imageName }}"
           imagePullPolicy: {{ .Values.image.pullPolicy }}
+          {{- with .Values.volumeMounts }}
           volumeMounts:
-            - mountPath: /models
-              name: models
+          {{- toYaml . | nindent 12 }}
+          {{- end }}
 
-          resources:
-            limits:
-              nvidia.com/gpu: {{ .Values.image.numGpus }}
+          {{- with .Values.envFrom }}
+          envFrom:
+          {{- toYaml . | nindent 12 }}
+          {{- end }}
 
+          {{- with .Values.resources }}
+          resources:
+          {{- toYaml . | nindent 12 }}
+          {{- end }}
           args:
             - tritonserver
             {{- range .Values.serverArgs }}

deploy/k8s-onprem/values.yaml

@@ -32,14 +32,31 @@ tags:
 image:
   imageName: nvcr.io/nvidia/tritonserver:25.07-py3
   pullPolicy: IfNotPresent
-  modelRepositoryServer: < Replace with the IP Address of your file server >
-  modelRepositoryPath: /srv/models
-  numGpus: 1
 
 # add server args here e.g. --grpc-use-ssl, --grpc-server-certs, repository-poll-secs, etc
 serverArgs:
   - '--model-repository=/models'
 
+envFrom: []
+# - secretRef:
+#     name: triton-secret-name
+
+volumes: 
+  - name: models
+    nfs:
+      server: < Replace with the IP Address of your file server >
+      path: /srv/models
+      readOnly: false
+
+volumeMounts:
+  - name: models
+    mountPath: /models
+
+# CPU and memory constraint can also be specified here
+resources: 
+  limits:
+    nvidia.com/gpu: 1
+
 traefik:
   ports:
     triton-http: