set probes

clusters/k3s-dgx/nim-service/qwen.yaml

@@ -52,7 +52,9 @@ spec:
     repository: vllm/vllm-openai
     tag: v0.19.1-cu130
     pullPolicy: IfNotPresent
-
+  env:
+    - name: VLLM_CACHE_ROOT
+      value: /model-store/vllm-cache
   command:
     - python3
   args:
@@ -91,6 +93,32 @@ spec:
   resources:
     limits:
       nvidia.com/gpu: 1
+  livenessProbe:
+    httpGet:
+      path: /health
+      port: 8000
+    initialDelaySeconds: 120
+    periodSeconds: 30
+    timeoutSeconds: 10
+    failureThreshold: 10
+
+  readinessProbe:
+    httpGet:
+      path: /health
+      port: 8000
+    initialDelaySeconds: 30
+    periodSeconds: 15
+    timeoutSeconds: 10
+    failureThreshold: 20
+
+  startupProbe:
+    httpGet:
+      path: /health
+      port: 8000
+    initialDelaySeconds: 10
+    periodSeconds: 20
+    timeoutSeconds: 10
+    failureThreshold: 60
   expose:
     service:
       type: ClusterIP