# SPDX-FileCopyrightText: Copyright (c) 2025 NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 apiVersion: nvidia.com/v1alpha1 kind: DynamoGraphDeployment metadata: name: trtllm-v1-disagg-router spec: services: Frontend: dynamoNamespace: trtllm-v1-disagg-router componentType: frontend replicas: 1 extraPodSpec: mainContainer: image: my-registry/trtllm-runtime:my-tag envs: - name: DYN_ROUTER_MODE value: kv TRTLLMPrefillWorker: dynamoNamespace: trtllm-v1-disagg-router envFromSecret: hf-token-secret componentType: worker replicas: 2 resources: limits: gpu: "1" extraPodSpec: mainContainer: image: my-registry/trtllm-runtime:my-tag workingDir: /workspace/ command: - python3 - -m - dynamo.trtllm args: - --model-path - Qwen/Qwen3-0.6B - --served-model-name - Qwen/Qwen3-0.6B - --extra-engine-args - ./recipes/qwen3/trtllm/prefill.yaml - --disaggregation-mode - prefill - --disaggregation-strategy - prefill_first - --publish-events-and-metrics TRTLLMDecodeWorker: dynamoNamespace: trtllm-v1-disagg-router envFromSecret: hf-token-secret componentType: worker replicas: 1 resources: limits: gpu: "1" extraPodSpec: mainContainer: image: my-registry/trtllm-runtime:my-tag workingDir: /workspace/ command: - python3 - -m - dynamo.trtllm args: - --model-path - Qwen/Qwen3-0.6B - --served-model-name - Qwen/Qwen3-0.6B - --extra-engine-args - ./recipes/qwen3/trtllm/decode.yaml - --disaggregation-mode - decode - --disaggregation-strategy - prefill_first