openai-api-route.yaml

apiVersion: apps/v1
kind: Deployment
metadata:
  name: openai-api-route
  labels:
    app: openai-api-route
spec:
  replicas: 2
  selector:
    matchLabels:
      app: openai-api-route
  template:
    metadata:
      labels:
        app: openai-api-route
    spec:
      imagePullSecrets:
      - name: spiderman-regcred
      volumes:
      - name: config-volume
        configMap:
          name: openai-api-route-config
      containers:
      - name: openai-api-route
        image: registry.waykey.net:7999/spiderman/datamining/openai-api-route:latest
        imagePullPolicy: Always
        ports:
        - containerPort: 8888
        resources:
          requests:
            memory: "400M"
          limits:
            memory: "1000M"
        volumeMounts:
        - name: config-volume
          mountPath: /config.yaml
          subPath:config.yaml
---
apiVersion: v1
kind: Service
metadata:
  labels:
    app: openai-api-route
  name: openai-api-route
  namespace: default
spec:
  ports:
  - port: 8888
    protocol: TCP
    targetPort: 8888
  selector:
    app: openai-api-route
---
apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
  name: openai-api-route
spec:
  ingressClassName: nginx
  rules:
  - host: openai-api-route.k8s.local
    http:
      paths:
      - path: /
        pathType: Prefix
        backend:
          service:
            name: openai-api-route
            port:
              number: 8888
---
apiVersion: v1
kind: ConfigMap
metadata:
  name: openai-api-route-config
  namespace: default
data:
  config.yaml: |
    authorization: n

    #dbtype: sqlite
    #dbaddr: /data/db.sqlite

    # 使用 postgres 作为数据库储存请求记录
    dbtype: postgres
    dbaddr: "host=192.168.1.20 port=5432 user=waykey dbname=openai_api_route sslmode=disable password=wk24edcRFV43321"

    upstreams:
      - endpoint: https://api.openai.com/v1
        sk: YOUR_API_SECRET_KEY
        timeout: 20
        deny:
          - gpt-4-32k
          - gpt-4-32k-1106