gcloud-lab/apps/base/customer1/hermes-agent/deployment.yaml

apiVersion: apps/v1
kind: Deployment
metadata:
  name: hermes-agent
  namespace: customer1
spec:
  replicas: 1
  selector:
    matchLabels:
      app: hermes-agent
  template:
    metadata:
      labels:
        app: hermes-agent
    spec:
      containers:
      - name: hermes-agent
        image: nousresearch/hermes-agent:latest
        command: ["/bin/bash", "-c"]
        args:
          - |
            # Make `hermes` CLI instantly available when you exec in
            ln -sf /opt/hermes/.venv/bin/hermes /usr/local/bin/hermes 2>/dev/null || true
            echo "Hermes Agent starting (Telegram polling + full config persistence)..."
            exec hermes gateway run
        env:
          # === Hermes Home - REQUIRED for config persistence ===
          - name: HERMES_HOME
            value: "/opt/data"

          # === Telegram Configuration (polling only - no public exposure) ===
          - name: TELEGRAM_BOT_TOKEN
            value: "YOUR_TELEGRAM_BOT_TOKEN_HERE"          # ← from @BotFather
          - name: TELEGRAM_ALLOWED_USERS
            value: "YOUR_TELEGRAM_USER_ID_HERE"            # ← numeric ID from @userinfobot

          # === Local vLLM (OpenAI-compatible) ===
          - name: OPENAI_BASE_URL
            value: "http://vllm-service.customer1.svc.cluster.local:8000/v1"   # ← adjust if your service name differs
          - name: OPENAI_API_KEY
            value: "dummy"                                 # vLLM ignores this

          # === Optional ===
          # - name: LOG_LEVEL
          #   value: "INFO"
        volumeMounts:
        - name: hermes-data
          mountPath: /opt/data
      volumes:
      - name: hermes-data
        persistentVolumeClaim:
          claimName: hermes-agent-pvc
---
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: hermes-agent-pvc
  namespace: customer1
spec:
  accessModes:
    - ReadWriteOnce
  resources:
    requests:
      storage: 5Gi