arc53--docsgpt/deployment/k8s/deployments/docsgpt-deploy.yaml

apiVersion: apps/v1
kind: Deployment
metadata:
  name: docsgpt-api
spec:
  replicas: 1
  selector:
    matchLabels:
      app: docsgpt-api
  template:
    metadata:
      labels:
        app: docsgpt-api
    spec:
      initContainers:
      # Block pod start until Postgres accepts connections. The `postgres-init`
      # Job is responsible for running alembic migrations; this container only
      # waits for the server to be reachable.
      - name: wait-for-postgres
        image: postgres:16-alpine
        command:
          - sh
          - -c
          - |
            until pg_isready -h postgres -p 5432 -U docsgpt -d docsgpt; do
              echo "Waiting for postgres..."; sleep 2;
            done
      containers:
      - name: docsgpt-api
        image: arc53/docsgpt
        ports:
        - containerPort: 7091
        resources:
          limits:
            memory: "4Gi"
            cpu: "2"
          requests:
            memory: "2Gi"
            cpu: "1"
        envFrom:
        - secretRef:
            name: docsgpt-secrets
        env:
        - name: FLASK_APP
          value: "application/app.py"
        - name: DEPLOYMENT_TYPE
          value: "cloud"
        - name: POSTGRES_URI
          valueFrom:
            secretKeyRef:
              name: docsgpt-secrets
              key: POSTGRES_URI
        # Disable in-app auto-bootstrap. The `postgres-init` Job under
        # deployment/k8s/jobs/ owns schema creation and Alembic migrations,
        # so application pods must not race with it on rollout.
        - name: AUTO_MIGRATE
          value: "false"
        - name: AUTO_CREATE_DB
          value: "false"
---
apiVersion: apps/v1
kind: Deployment
metadata:
  name: docsgpt-worker
spec:
  replicas: 1
  selector:
    matchLabels:
      app: docsgpt-worker
  template:
    metadata:
      labels:
        app: docsgpt-worker
    spec:
      initContainers:
      - name: wait-for-postgres
        image: postgres:16-alpine
        command:
          - sh
          - -c
          - |
            until pg_isready -h postgres -p 5432 -U docsgpt -d docsgpt; do
              echo "Waiting for postgres..."; sleep 2;
            done
      containers:
      - name: docsgpt-worker
        image: arc53/docsgpt
        # `parsing` queue carries read_document/parse_document; required for its await to resolve.
        # For heavy/OCR parsing, run a separate deployment with `-Q parsing` (and GPU env).
        command: ["celery", "-A", "application.app.celery", "worker", "-l", "INFO", "-n", "worker.%h", "-Q", "docsgpt,parsing"]
        resources:
          limits:
            memory: "4Gi"
            cpu: "2"
          requests:
            memory: "2Gi"
            cpu: "1"
        envFrom:
        - secretRef:
            name: docsgpt-secrets
        env:
        - name: API_URL
          value: "http://<your-api-endpoint>"
        - name: POSTGRES_URI
          valueFrom:
            secretKeyRef:
              name: docsgpt-secrets
              key: POSTGRES_URI
        # Disable in-app auto-bootstrap. The `postgres-init` Job under
        # deployment/k8s/jobs/ owns schema creation and Alembic migrations,
        # so application pods must not race with it on rollout.
        - name: AUTO_MIGRATE
          value: "false"
        - name: AUTO_CREATE_DB
          value: "false"
---
apiVersion: apps/v1
kind: Deployment
metadata:
  name: docsgpt-frontend
spec:
  replicas: 1
  selector:
    matchLabels:
      app: docsgpt-frontend
  template:
    metadata:
      labels:
        app: docsgpt-frontend
    spec:
      containers:
      - name: docsgpt-frontend
        image: arc53/docsgpt-fe
        ports:
        - containerPort: 5173
        resources:
          limits:
            memory: "1Gi"
            cpu: "1"
          requests:
            memory: "256Mi"
            cpu: "100m"
        env:
        - name: VITE_API_HOST
          value: "http://<your-api-endpoint>"
        - name: VITE_API_STREAMING
          value: "true"