lumen/docker-compose.yml.example at main · ncsa/lumen · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
services:
  traefik:
    image: "traefik:3"
    container_name: traefik
    command:
      - --log.level=INFO
      - --api=true
      - --api.dashboard=true
      - --api.insecure=true
      # Entrypoints
      - --entrypoints.http.address=:80
      - --entrypoints.http.http.redirections.entryPoint.to=https
      - --entrypoints.https.address=:443
      - --entrypoints.https.http.tls.certresolver=myresolver
      # letsencrypt
      - --certificatesresolvers.myresolver.acme.email=admin@example.com
      - --certificatesresolvers.myresolver.acme.storage=/config/acme.json
      # uncomment to use testing certs
      - --certificatesresolvers.myresolver.acme.httpchallenge=true
      - --certificatesresolvers.myresolver.acme.httpchallenge.entrypoint=http
      # Docker setup
      - --providers.docker=true
      - --providers.docker.endpoint=unix:///var/run/docker.sock
      - --providers.docker.exposedbydefault=false
      - --providers.docker.watch=true
    restart: "unless-stopped"
    security_opt:
      - no-new-privileges:true
    ports:
      - "80:80"
      - "443:443"
      - "127.0.0.1:8080:8080"
    volumes:
      - "./traefik:/config"
      - "/var/run/docker.sock:/var/run/docker.sock:ro"

  postgres:
    image: timescale/timescaledb:latest-pg17
    container_name: postgres
    restart: unless-stopped
    environment:
      - POSTGRES_DB=lumen
      - POSTGRES_USER=lumen
      - POSTGRES_PASSWORD=lumen
    volumes:
      - postgres_data:/var/lib/postgresql/data
    healthcheck:
      test: ["CMD-SHELL", "pg_isready -U lumen"]
      interval: 10s
      timeout: 5s
      retries: 5

  # frontend
  lumen:
    image: ncsa/lumen:latest
    container_name: lumen
    restart: unless-stopped
    depends_on:
      postgres:
        condition: service_healthy
    environment:
      - CONFIG_YAML=/app/instance/config.yaml
      - DATABASE_URL=postgresql://lumen:lumen@postgres/lumen
    volumes:
      - ./lumen:/app/instance
    labels:
      - "traefik.enable=true"
      - "traefik.http.routers.lumen.entrypoints=https"
      - "traefik.http.routers.lumen.rule=Host(`lumen.example.com`)"

  # models
  # make sure to add to lumen/config.yaml
  nemotron-cascade-2:
    image: vllm/vllm-openai:v0.18.0
    container_name: nemotron-cascade-2
    command: >
      --port 8000
      --enable-auto-tool-choice
      --tool-call-parser hermes
      --gpu-memory-utilization 0.93
      --model nvidia/Nemotron-Cascade-2-30B-A3B
      --max-model-len 32768
      --max-num-seqs 32
      --attention-backend flashinfer
      --enable-prefix-caching
      --trust-remote-code
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              device_ids: ["0"]
              capabilities: [gpu]
    shm_size: "16gb"
    healthcheck:
      test: ["CMD", "curl", "-f", "http://localhost:8000/health"]
      interval: 30s
      timeout: 10s
      retries: 5
      start_period: 600s
    environment:
      - HF_HUB_OFFLINE=1 # skip huggingface.co network calls entirely
      - HF_DATASETS_OFFLINE=1
      - TRANSFORMERS_OFFLINE=1
      - TRITON_CACHE_DIR=/root/.triton/cache
      - CUDA_CACHE_PATH=/root/.nv/ComputeCache
      - CUDA_CACHE_MAXSIZE=10737418240
    volumes:
      - ./cache:/root/.cache/huggingface
      - ./vllm_cache:/root/.cache/vllm
      - ./triton_cache:/root/.triton/cache

  qwen3-coder-next:
    image: nvcr.io/nvidia/vllm:26.01-py3
    container_name: qwen3-coder-next
    command: >
      vllm serve Qwen/Qwen3-Coder-Next-FP8
      --port 8001
      --enable-auto-tool-choice
      --tool-call-parser qwen3_coder
      --gpu-memory-utilization 0.8
      --max-model-len 131072
      --attention-backend flashinfer
      --enable-prefix-caching
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              device_ids: ["1"]
              capabilities: [gpu]
    shm_size: "16gb"
    healthcheck:
      test: ["CMD", "curl", "-f", "http://localhost:8001/health"]
      interval: 30s
      timeout: 10s
      retries: 5
      start_period: 600s
    environment:
      - HF_HUB_OFFLINE=1 # skip huggingface.co network calls entirely
      - HF_DATASETS_OFFLINE=1
      - TRANSFORMERS_OFFLINE=1
      - TRITON_CACHE_DIR=/root/.triton/cache
      - CUDA_CACHE_PATH=/root/.nv/ComputeCache
      - CUDA_CACHE_MAXSIZE=10737418240
    volumes:
      - ./cache:/root/.cache/huggingface
      - ./vllm_cache:/root/.cache/vllm
      - ./triton_cache:/root/.triton/cache

volumes:
  postgres_data: