BloomFLv2/docker-compose.yml at main · AdvaitVarhade/BloomFLv2 · 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
# BloomFL — docker-compose deployment
#
# Scalable decentralised edge-intelligence nodes.
# Each replica runs on the host network (required for mDNS multicast).
#
# Quick start (single node):
#   docker-compose up --build
#
# Scale to 10 nodes (each gets a unique port via entrypoint.sh):
#   docker-compose up --build --scale node=10
#
# Override transport or gossip interval per deployment:
#   BLOOMFL_TRANSPORT=grpc docker-compose up --scale node=5

services:
  node:
    build:
      context: .
      dockerfile: Dockerfile
    image: bloomfl:latest

    # host networking is required on Linux so mDNS multicast (224.0.0.251)
    # works across all replicas on the same machine.
    network_mode: host

    restart: unless-stopped

    environment:
      # Core node settings — override as needed
      BLOOMFL_TRANSPORT: ${BLOOMFL_TRANSPORT:-tcp}
      BLOOMFL_GOSSIP_INTERVAL_SECONDS: ${BLOOMFL_GOSSIP_INTERVAL_SECONDS:-15}
      BLOOMFL_TRAIN_EPOCHS_PER_ROUND: ${BLOOMFL_TRAIN_EPOCHS_PER_ROUND:-1}
      BLOOMFL_ADAPTATION_ENABLED: ${BLOOMFL_ADAPTATION_ENABLED:-true}
      BLOOMFL_AGGREGATION_STRATEGY: ${BLOOMFL_AGGREGATION_STRATEGY:-weighted_avg}
      # Data / storage inside container home
      BLOOMFL_KEY_STORAGE_DIR: /home/bloomfl/.bloomfl/keys
      BLOOMFL_METRICS_DIR: /home/bloomfl/.bloomfl/metrics
      BLOOMFL_DATA_DIR: /home/bloomfl/.bloomfl/data
      # Optional: static peer list (comma-separated host:port) if mDNS unavailable
      BLOOMFL_PEER_ADDRS: ${BLOOMFL_PEER_ADDRS:-}

    volumes:
      # Mount local results directory for metric persistence
      - ./results/metrics:/home/bloomfl/.bloomfl/metrics
      - ./results/keys:/home/bloomfl/.bloomfl/keys

    # Resource limits — reasonable defaults for edge devices
    deploy:
      replicas: 1     # override with --scale node=N
      resources:
        limits:
          cpus: "1.0"
          memory: 1G
        reservations:
          cpus: "0.25"
          memory: 256M

    logging:
      driver: "json-file"
      options:
        max-size: "50m"
        max-file: "3"

    healthcheck:
      test: ["CMD", "nc", "-z", "localhost", "${BLOOMFL_LISTEN_PORT:-50051}"]
      interval: 30s
      timeout: 5s
      retries: 3
      start_period: 30s

  # ── Dashboard API (FastAPI) ────────────────────────────────────────────────
  api:
    build:
      context: .
      dockerfile: Dockerfile
    image: bloomfl:latest
    command: >
      sh -c "/home/bloomfl/.bloomfl/venv/bin/uvicorn api.main:app
             --host 0.0.0.0 --port 8000 --log-level info"
    ports:
      - "8000:8000"
    environment:
      BLOOMFL_METRICS_DIR: /home/bloomfl/.bloomfl/metrics
      BLOOMFL_KEY_STORAGE_DIR: /home/bloomfl/.bloomfl/keys
      BLOOMFL_DATA_DIR: /home/bloomfl/.bloomfl/data
    volumes:
      - ./results/metrics:/home/bloomfl/.bloomfl/metrics
      - ./results/keys:/home/bloomfl/.bloomfl/keys
      - .:/app
    working_dir: /app
    depends_on:
      - node
    healthcheck:
      test: ["CMD", "curl", "-f", "http://localhost:8000/health"]
      interval: 15s
      timeout: 5s
      retries: 3
      start_period: 10s
    logging:
      driver: "json-file"
      options:
        max-size: "20m"
        max-file: "2"

  # ── Dashboard Frontend (Next.js) ───────────────────────────────────────────
  frontend:
    image: node:20-alpine
    working_dir: /app
    command: sh -c "npm run build && npm start"
    ports:
      - "3000:3000"
    environment:
      NEXT_PUBLIC_API_URL: ${NEXT_PUBLIC_API_URL:-http://api:8000}
      NODE_ENV: production
      PORT: 3000
    volumes:
      - ./frontend:/app
      - /app/node_modules
    depends_on:
      - api
    healthcheck:
      test: ["CMD", "wget", "-qO-", "http://localhost:3000"]
      interval: 30s
      timeout: 5s
      retries: 3
      start_period: 60s
    logging:
      driver: "json-file"
      options:
        max-size: "20m"
        max-file: "2"