aibrix/config/overlays/vke/default/envoy_proxy_patch.yaml at 09a7fa16a73ec7c0ab068cebaf2e077d056991b7 · vllm-project/aibrix · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
apiVersion: gateway.envoyproxy.io/v1alpha1
kind: EnvoyProxy
metadata:
  name: aibrix-custom-proxy-config
  namespace: aibrix-system
spec:
  provider:
    type: Kubernetes
    kubernetes:
      envoyService:
        patch:
          type: StrategicMerge
          value:
            spec:
              ipFamilies:
                - IPv4
              ipFamilyPolicy: SingleStack
              type: LoadBalancer
      envoyDeployment:
        replicas: 1
        pod:
          affinity:
            nodeAffinity: # prevent gateway pod to be placed on gpu node.
              preferredDuringSchedulingIgnoredDuringExecution:
              - weight: 100
                preference:
                  matchExpressions:
                    - key: vke.node.gpu.schedule
                      operator: NotIn
                      values:
                        - nvidia
        patch:
          type: StrategicMerge
          value:
            spec:
              template:
                spec:
                  containers:
                    - name: envoy
                      image: aibrix-cn-beijing.cr.volces.com/aibrix/envoy:v1.33.2
                      resources:
                        requests:
                          cpu: 2
                          memory: 8Gi
                        limits:
                          cpu: 2
                          memory: 8Gi
                    - name: shutdown-manager
                      image: aibrix-cn-beijing.cr.volces.com/aibrix/gateway:v1.6.3
                      resources:
                        requests:
                          cpu: 10m
                          memory: 32Mi