docker-compose.yml

services:

  kafka:
    image: apache/kafka:3.9.0
    ports:
      - "9092:9092"
    environment:
      KAFKA_NODE_ID: "1"
      KAFKA_PROCESS_ROLES: "broker,controller"
      KAFKA_CONTROLLER_QUORUM_VOTERS: "1@kafka:9093"
      KAFKA_LISTENERS: "PLAINTEXT://:9092,CONTROLLER://:9093"
      KAFKA_ADVERTISED_LISTENERS: "PLAINTEXT://kafka:9092"
      KAFKA_LISTENER_SECURITY_PROTOCOL_MAP: "CONTROLLER:PLAINTEXT,PLAINTEXT:PLAINTEXT"
      KAFKA_CONTROLLER_LISTENER_NAMES: "CONTROLLER"
      KAFKA_INTER_BROKER_LISTENER_NAME: "PLAINTEXT"
      CLUSTER_ID: "MkU3OEVBNTcwNTJENDM2Qk"
      KAFKA_AUTO_CREATE_TOPICS_ENABLE: "true"
      KAFKA_NUM_PARTITIONS: "3"
    healthcheck:
      test: ["CMD-SHELL", "/opt/kafka/bin/kafka-topics.sh --bootstrap-server localhost:9092 --list"]
      interval: 10s
      timeout: 5s
      retries: 15
    deploy:
      resources:
        limits:
          memory: 1G

  mlflow:
    image: python:3.12-slim
    command: >
      bash -c "pip install mlflow>=2.13 --quiet --no-cache-dir &&
      mlflow server --backend-store-uri file:///mlruns --default-artifact-root file:///mlruns --host 0.0.0.0 --port 5000"
    ports:
      - "5000:5000"
    volumes:
      - ./data/mlruns:/mlruns

  spark:
    build: ./services/spark-jobs
    command: >
      bash -c "until [ -f /data/ml-1m/ratings.dat ]; do echo 'Waiting for bootstrap data...'; sleep 5; done && python streaming.py"
    environment:
      SPARK_MASTER: "local[*]"
      REDIS_URL: "redis://redis:6379"
      POSTGRES_URL: "postgresql://recsys:recsys@postgres:5432/recsys"
      KAFKA_BOOTSTRAP_SERVERS: "kafka:9092"
      DATA_DIR: "/data"
      CHECKPOINT_DIR: "/tmp/spark-checkpoint"
      MLFLOW_TRACKING_URI: "http://mlflow:5000"
      PYTHONUNBUFFERED: "1"
    volumes:
      - ./data:/data
      - spark-checkpoint:/tmp/spark-checkpoint
    depends_on:
      kafka:   { condition: service_healthy }
      redis:   { condition: service_healthy }
      postgres: { condition: service_healthy }
    deploy:
      resources:
        limits:
          memory: 2G
    restart: on-failure

  event-simulator:
    build: ./services/event-simulator
    environment:
      KAFKA_BOOTSTRAP_SERVERS: "kafka:9092"
      DATA_DIR: "/data"
      EVENT_RATE: "5"
    volumes:
      - ./data:/data
    depends_on:
      kafka: { condition: service_healthy }
    restart: on-failure

  api:
    build: ./services/api
    ports:
      - "8000:8000"
    environment:
      REDIS_URL: "redis://redis:6379"
      POSTGRES_URL: "postgresql://recsys:recsys@postgres:5432/recsys"
      KAFKA_BOOTSTRAP_SERVERS: "kafka:9092"
    depends_on:
      redis:    { condition: service_healthy }
      postgres: { condition: service_healthy }
      kafka:    { condition: service_healthy }
    healthcheck:
      test: ["CMD-SHELL", "python3 -c \"import urllib.request; urllib.request.urlopen('http://localhost:8000/health')\""]
      interval: 10s
      timeout: 5s
      retries: 5
    restart: on-failure

  redis:
    image: redis:7-alpine
    ports:
      - "6379:6379"
    command: redis-server --save 60 1 --appendonly yes
    volumes:
      - redis-data:/data
    healthcheck:
      test: ["CMD", "redis-cli", "ping"]
      interval: 5s
      timeout: 3s
      retries: 10

  postgres:
    image: postgres:16-alpine
    environment:
      POSTGRES_USER: ${POSTGRES_USER:-recsys}
      POSTGRES_PASSWORD: ${POSTGRES_PASSWORD:-recsys}
      POSTGRES_DB: ${POSTGRES_DB:-recsys}
    ports:
      - "5432:5432"
    volumes:
      - postgres-data:/var/lib/postgresql/data
      - ./services/api/schema.sql:/docker-entrypoint-initdb.d/schema.sql
    healthcheck:
      test: ["CMD-SHELL", "pg_isready -U recsys"]
      interval: 5s
      timeout: 3s
      retries: 10

  frontend:
    build: ./services/frontend
    ports:
      - "3000:80"
    depends_on:
      api: { condition: service_healthy }

  prometheus:
    image: prom/prometheus:v2.51.0
    ports:
      - "9090:9090"
    volumes:
      - ./infra/prometheus.yml:/etc/prometheus/prometheus.yml:ro

  grafana:
    image: grafana/grafana:10.4.3
    ports:
      - "3001:3000"
    environment:
      GF_SECURITY_ADMIN_PASSWORD: ${GRAFANA_ADMIN_PASSWORD:-admin}
      GF_AUTH_ANONYMOUS_ENABLED: "true"
      GF_AUTH_ANONYMOUS_ORG_ROLE: Viewer
    volumes:
      - grafana-data:/var/lib/grafana
      - ./infra/grafana/datasources:/etc/grafana/provisioning/datasources:ro
      - ./infra/grafana/dashboards:/etc/grafana/provisioning/dashboards:ro

volumes:
  postgres-data:
  grafana-data:
  spark-checkpoint:
  redis-data: