Merge pull request #64 from sirius0xdev/feat/optimize-rtx6000-vllm-keda-scaling

feat: optimize RTX6000 vLLM KEDA scaling

infrastructure/gpus/base/keda-gpu-scaling/keda-vllm.yaml

@@ -1,17 +1,22 @@
-apiVersion: keda.sh/v1alpha1
-kind: ScaledObject
+apiVersion: http.keda.sh/v1alpha1
+kind: HTTPScaledObject
 metadata:
   name: rtx6000-scaling
   namespace: customer1
 spec:
+  hosts:
+  - rtx6000-brain-service.customer1.svc.cluster.local
   scaleTargetRef:
     name: rtx6000-brain-vllm
     kind: Deployment
     apiVersion: apps/v1
-  hosts:
-    - rtx6000-brain-service.customer1.svc.cluster.local
+    service: rtx6000-brain-service
+    port: 8000
+  paths:
+  - path: "^/v1/(chat/)?completions"
   replicas:
     min: 0
     max: 1
-  targetPendingRequests: 1
+  pollingInterval: 10
+  cooldownPeriod: 30
   scaledownPeriod: 900

infrastructure/gpus/base/vllm-servers/rtx6000-vllm.yaml

@@ -2,12 +2,12 @@
 apiVersion: apps/v1
 kind: Deployment
 metadata:
-  name: openclaw-brain-vllm
+  name: rtx6000-brain-vllm
   namespace: customer1
   labels:
     app: rtx6000-brain
 spec:
-  replicas: 1
+  replicas: 0
   selector:
     matchLabels:
       app: rtx6000-brain