agg.yaml

# SPDX-FileCopyrightText: Copyright (c) 2025-2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
# SPDX-License-Identifier: Apache-2.0

# NOTE: There is no dedicated `mocker-runtime` image. The mocker component
# (`python3 -m dynamo.mocker`) is bundled into the standard Dynamo backend
# runtime images published on NGC (e.g. `nvcr.io/nvidia/ai-dynamo/vllm-runtime`,
# `sglang-runtime`, `tensorrtllm-runtime`). Replace `my-dynamo-image:my-tag`
# below with any of those tags, or with your own built image.
apiVersion: nvidia.com/v1alpha1
kind: DynamoGraphDeployment
metadata:
  name: mocker-agg
spec:
  services:
    Frontend:
      componentType: frontend
      replicas: 1
      extraPodSpec:
        mainContainer:
          image: my-dynamo-image:my-tag
    decode:
      envFromSecret: hf-token-secret
      componentType: worker
      subComponentType: decode
      replicas: 1
      extraPodSpec:
        mainContainer:
          image: my-dynamo-image:my-tag
          workingDir: /workspace
          command:
          - python3
          - -m
          - dynamo.mocker
          args:
            - --model-path
            - nvidia/Llama-3.1-8B-Instruct-FP8
            - --model-name
            - nvidia/Llama-3.1-8B-Instruct-FP8
            - --speedup-ratio
            - "1.0"
            - --planner-profile-data
            - /workspace/components/src/dynamo/planner/tests/data/profiling_results/H200_TP1P_TP1D