version: '3.7'

services:

  # A proxy based on nginx that sits between the Plex server and
  # the internet. Every time a request is made to Plex, if that
  # request is to mark status, an API call is made
  # to plex-cluster-manager.
  #
  # There is one of these for every Plex server.
  plex-cluster-proxy:
    image: plex-cluster-proxy
    environment:
      - PLEX_IP=192.168.6.200
      - PLEX_PORT=32400
      - HTTPS_HOST=plex-northwyck.nowsci.com
      - HTTPS_PORT=33400
      - RESOLVERS=8.8.4.4 8.8.8.8
      - SSL_CERTIFICATE=/certs/fullchain.pem
      - SSL_CERTIFICATE_KEY=/certs/privkey.pem
      - CLUSTER_MANAGER=http://plex-cluster-manager:3400
      - CLUSTER_MANAGER_TOKEN=ts3wJgUNaJ6DcB7FhhYGcUpaa9DwPy
    volumes:
      - /etc/localtime:/etc/localtime:ro
      - ./certs/fullchain.pem:/certs/fullchain.pem:ro
      - ./certs/privkey.pem:/certs/privkey.pem:ro
      - plex-cluster-proxy:/data
    ports:
      - 33400:33400
    restart: always

  # The service that sychronizes the Plex servers. It takes the API calls
  # from plex-cluster-proxy, and reaches out to any other configured Plex
  # servers to set their status just as if the client connecting to the
  # original Plex server was connecting to that one instead.
  #
  # You will want to secure this behind an SSL proxy in the real world.
  #
  # There is only ever one of these.
  plex-cluster-manager:
    image: plex-cluster-manager
    command: 'npm run dev'
    environment:
      - CLUSTER_MANAGER_TOKEN=ts3wJgUNaJ6DcB7FhhYGcUpaa9DwPy
    volumes:
      - /etc/localtime:/etc/localtime:ro
      - ./config/plex-cluster-manager.yml:/config/plex-cluster-manager.yml:ro
      - ./plex-cluster-manager/src/:/app:z
      - /app/node_modules
      - plex-cluster-manager:/data
    restart: always

volumes:
  plex-cluster-manager:
  plex-cluster-proxy: