vars:
  clusterName:
    default: "inference-sim"
    description: "Name of the Kind cluster"
  imageRepository:
    default: "ghcr.io/llm-d/llm-d-inference-sim"
    description: "llm-d inference simulator Docker image repository"
  imageTag:
    default: "v0.10.2"
    description: "llm-d inference simulator Docker image tag"
  model:
    default: "gpt-4o"
    description: "Model name the simulator loads and serves"
  mode:
    default: "echo"
    description: "Simulator mode: echo returns the request text back, random returns canned sentences"
  seed:
    default: "42"
    description: "Random seed, keeps generated responses reproducible"
images:
  preload:
    refs:
      - "{{ .imageRepository }}:{{ .imageTag }}"
kind:
  name: "{{ .clusterName }}"
  nodes:
    - role: control-plane
      extraPortMappings:
        - containerPort: 30802
          hostPort: 30802
components:
  - name: inference-sim
    type: k8s
    k8s:
      manifests:
        - apiVersion: apps/v1
          kind: Deployment
          metadata:
            name: inference-sim
            labels:
              app: inference-sim
          spec:
            replicas: 1
            selector:
              matchLabels:
                app: inference-sim
            template:
              metadata:
                labels:
                  app: inference-sim
              spec:
                containers:
                  - name: inference-sim
                    image: "{{ .imageRepository }}:{{ .imageTag }}"
                    args:
                      - --model={{ .model }}
                      - --mode={{ .mode }}
                      - --seed={{ .seed }}
                      - --port=8000
                    ports:
                      - containerPort: 8000
                        name: http
                    readinessProbe:
                      httpGet:
                        path: /ready
                        port: 8000
                    livenessProbe:
                      httpGet:
                        path: /health
                        port: 8000
                    resources:
                      requests:
                        cpu: 50m
                        memory: 64Mi
                      limits:
                        cpu: 200m
                        memory: 256Mi
        - apiVersion: v1
          kind: Service
          metadata:
            name: inference-sim
            labels:
              app: inference-sim
          spec:
            type: NodePort
            ports:
              - port: 8000
                targetPort: 8000
                nodePort: 30802
                name: http
            selector:
              app: inference-sim