disagg.yaml

# 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: vllm-disagg
spec:
  services:
    Frontend:
      dynamoNamespace: vllm-disagg
      componentType: main
      replicas: 1
      livenessProbe:
        httpGet:
          path: /health
          port: 8000
        initialDelaySeconds: 60
        periodSeconds: 60
        timeoutSeconds: 30
        failureThreshold: 10
      readinessProbe:
        exec:
          command:
            - /bin/sh
            - -c
            - 'curl -s http://localhost:8000/health | jq -e ".status == \"healthy\""'
        initialDelaySeconds: 60
        periodSeconds: 60
        timeoutSeconds: 30
        failureThreshold: 10
      resources:
        requests:
          cpu: "32"
          memory: "10Gi"
        limits:
          cpu: "32"
          memory: "10Gi"
      extraPodSpec:
        mainContainer:
          image: nvcr.io/nvidian/nim-llm-dev/vllm-runtime:dep-233.17
          workingDir: /workspace/components/backends/vllm
          args:
            - "python3 -m dynamo.frontend --http-port 8000"
    VllmDecodeWorker:
      dynamoNamespace: vllm-disagg
      envFromSecret: hf-token-secret
      componentType: worker
      replicas: 1
      livenessProbe:
        exec:
          command:
            - /bin/sh
            - -c
            - "exit 0"
        periodSeconds: 60
        timeoutSeconds: 30
        failureThreshold: 10
      readinessProbe:
        exec:
          command:
            - /bin/sh
            - -c
            - 'grep "VllmWorker.*has been initialized" /tmp/vllm.log'
        initialDelaySeconds: 60
        periodSeconds: 60
        timeoutSeconds: 30
        failureThreshold: 10
      resources:
        requests:
          cpu: "32"
          memory: "40Gi"
          gpu: "1"
        limits:
          cpu: "32"
          memory: "40Gi"
          gpu: "1"
      extraPodSpec:
        mainContainer:
          image: nvcr.io/nvidian/nim-llm-dev/vllm-runtime:dep-233.17
          workingDir: /workspace/components/backends/vllm
          command:
            - /bin/sh
            - -c
          args:
            - "python3 -m dynamo.vllm --model Qwen/Qwen3-0.6B --enforce-eager 2>&1 | tee /tmp/vllm.log"
    VllmPrefillWorker:
      dynamoNamespace: vllm-disagg
      envFromSecret: hf-token-secret
      componentType: worker
      replicas: 1
      livenessProbe:
        exec:
          command:
            - /bin/sh
            - -c
            - "exit 0"
        periodSeconds: 60
        timeoutSeconds: 30
        failureThreshold: 10
      readinessProbe:
        exec:
          command:
            - /bin/sh
            - -c
            - 'grep "VllmWorker.*has been initialized" /tmp/vllm.log'
        initialDelaySeconds: 60
        periodSeconds: 60
        timeoutSeconds: 30
        failureThreshold: 10
      resources:
        requests:
          cpu: "32"
          memory: "40Gi"
          gpu: "1"
        limits:
          cpu: "32"
          memory: "40Gi"
          gpu: "1"
      extraPodSpec:
        mainContainer:
          image: nvcr.io/nvidian/nim-llm-dev/vllm-runtime:dep-233.17
          workingDir: /workspace/components/backends/vllm
          command:
            - /bin/sh
            - -c
          args:
            - "python3 -m dynamo.vllm --model Qwen/Qwen3-0.6B --enforce-eager --is-prefill-worker 2>&1 | tee /tmp/vllm.log"