LokiStack/docker-compose.prod.yml

# ============================================================
# opsx-loki - Production Docker Compose Configuration
# ============================================================
# Features:
#   - Multi-instance deployment with load balancing
#   - Read-only root filesystem for security
#   - tmpfs for temporary directories
#   - Stricter resource limits
#   - Docker Secrets support
#   - Health checks with longer intervals
#   - Rolling update configuration
# ============================================================

services:
  # ---- Nginx Load Balancer ----
  nginx:
    image: nginx:1.27-alpine
    container_name: opsxloki-lb
    restart: unless-stopped
    ports:
      - "8080:8080"
    volumes:
      - ./nginx.conf:/etc/nginx/nginx.conf:ro
    networks:
      - opsxloki-net
    healthcheck:
      test: ["CMD", "wget", "--no-verbose", "--tries=1", "--spider", "http://localhost:8080/health"]
      interval: 30s
      timeout: 5s
      retries: 3
    deploy:
      resources:
        limits:
          cpus: '0.5'
          memory: 128M
        reservations:
          cpus: '0.1'
          memory: 32M
    logging:
      driver: "json-file"
      options:
        max-size: "20m"
        max-file: "3"
    security_opt:
      - no-new-privileges:true
    cap_drop:
      - ALL
    cap_add:
      - NET_BIND_SERVICE
    read_only: true
    tmpfs:
      - /var/cache/nginx:noexec,nosuid,size=10m
      - /var/run:noexec,nosuid,size=5m
    depends_on:
      opsxloki-1:
        condition: service_healthy
      opsxloki-2:
        condition: service_healthy

  # ---- Application Instance 1 ----
  opsxloki-1:
    build:
      context: .
      dockerfile: Dockerfile
      args:
        TARGETARCH: ${TARGETARCH:-amd64}
    image: opsxloki:v2
    container_name: opsxloki-1
    hostname: opsxloki-1
    restart: unless-stopped

    security_opt:
      - no-new-privileges:true
    cap_drop:
      - ALL
    cap_add:
      - NET_BIND_SERVICE

    # P0: Read-only root filesystem
    read_only: true

    # P0: tmpfs for writable directories
    tmpfs:
      - /tmp:noexec,nosuid,size=50m,uid=1000,gid=1000

    volumes:
      # Named volume for data persistence
      - opsxloki-data-1:/app/data
      # kubeconfig read-only (P0: non-root path)
      - ${KUBECONFIG_PATH:-~/.kube}:/home/opsxloki/.kube:ro
      # SSH keys read-only
      - ${SSH_KEY_PATH:-~/.ssh}:/home/opsxloki/.ssh:ro

    environment:
      - SERVER_PORT=8080
      - DATA_DIR=/app/data
      - KUBECONFIG=/home/opsxloki/.kube/config
      # Secrets via Docker Secrets file injection
      - JWT_SECRET_FILE=${JWT_SECRET_FILE:-/run/secrets/jwt_secret}
      - ENCRYPTION_KEY_FILE=${ENCRYPTION_KEY_FILE:-/run/secrets/encryption_key}

    secrets:
      - jwt_secret
      - encryption_key

    # Health check
    healthcheck:
      test: ["CMD", "wget", "--no-verbose", "--tries=1", "--spider", "http://localhost:8080/index.html"]
      interval: 30s
      timeout: 5s
      retries: 3
      start_period: 20s

    # Stricter resource limits for production
    deploy:
      resources:
        limits:
          cpus: '1.0'
          memory: 256M
          pids: 50
        reservations:
          cpus: '0.25'
          memory: 64M
      restart_policy:
        condition: on-failure
        delay: 5s
        max_attempts: 3
        window: 60s

    logging:
      driver: "json-file"
      options:
        max-size: "50m"
        max-file: "5"
        labels: "service,instance"

    stop_grace_period: 30s
    networks:
      - opsxloki-net

  # ---- Application Instance 2 ----
  opsxloki-2:
    build:
      context: .
      dockerfile: Dockerfile
      args:
        TARGETARCH: ${TARGETARCH:-amd64}
    image: opsxloki:v2
    container_name: opsxloki-2
    hostname: opsxloki-2
    restart: unless-stopped

    security_opt:
      - no-new-privileges:true
    cap_drop:
      - ALL
    cap_add:
      - NET_BIND_SERVICE

    read_only: true

    tmpfs:
      - /tmp:noexec,nosuid,size=50m,uid=1000,gid=1000

    volumes:
      - opsxloki-data-2:/app/data
      - ${KUBECONFIG_PATH:-~/.kube}:/home/opsxloki/.kube:ro
      - ${SSH_KEY_PATH:-~/.ssh}:/home/opsxloki/.ssh:ro

    environment:
      - SERVER_PORT=8080
      - DATA_DIR=/app/data
      - KUBECONFIG=/home/opsxloki/.kube/config
      - JWT_SECRET_FILE=${JWT_SECRET_FILE:-/run/secrets/jwt_secret}
      - ENCRYPTION_KEY_FILE=${ENCRYPTION_KEY_FILE:-/run/secrets/encryption_key}

    secrets:
      - jwt_secret
      - encryption_key

    healthcheck:
      test: ["CMD", "wget", "--no-verbose", "--tries=1", "--spider", "http://localhost:8080/index.html"]
      interval: 30s
      timeout: 5s
      retries: 3
      start_period: 20s

    deploy:
      resources:
        limits:
          cpus: '1.0'
          memory: 256M
          pids: 50
        reservations:
          cpus: '0.25'
          memory: 64M
      restart_policy:
        condition: on-failure
        delay: 5s
        max_attempts: 3
        window: 60s

    logging:
      driver: "json-file"
      options:
        max-size: "50m"
        max-file: "5"
        labels: "service,instance"

    stop_grace_period: 30s
    networks:
      - opsxloki-net

# ---- Named Volumes ----
volumes:
  opsxloki-data-1:
    driver: local
  opsxloki-data-2:
    driver: local

# ---- Docker Secrets ----
secrets:
  jwt_secret:
    file: ${JWT_SECRET_FILE_PATH:-./secrets/jwt_secret.txt}
  encryption_key:
    file: ${ENCRYPTION_KEY_FILE_PATH:-./secrets/encryption_key.txt}

# ---- Network ----
networks:
  opsxloki-net:
    driver: bridge
    ipam:
      config:
        - subnet: 172.28.1.0/24