Migrate from swarm to dokploy: traefik stopper (single point of failure and remote ip pass through)

[devMls]

Hi

I am evaluating a migration from swarm.

I love the idea of CI/CD and database managment (today managed by cronjob docker image and mazzolino sheperd). the shared variables and secrets looks promising too.

But I am worry about the traefik reverse proxy.

Now I have a global caddy host mode with redis for shared state and certificates. This give me the chance of know the clients ip from the containers and a better HA .

with plugin https://github.com/lucaslorentz/caddy-docker-proxy I have autoconfig too.

Traefik in dokploy give me same features?

Can I replace with my caddy stack? I imagine that is possible but I will lose the "magic" of dokploy.

Here my current caddy stack:

version: "3.7"

services:
  # --- REDIS (Almacenamiento compartido para Caddy) ---
  redis_caddy:
    image: redis:alpine
    # Activamos persistencia AOF para no perder certificados en reinicios
    command: redis-server --appendonly yes
    networks:
      - network_general_internal
    volumes:
      - redis_data_caddy:/data
    deploy:
      replicas: 1
      restart_policy:
        condition: any
      placement:
        # Fijamos Redis al nodo con el disco persistente
        constraints: [node.labels.node==node1]

  # --- CADDY PROXY (Escalable y con Rate Limit) ---
  caddy:
    # Usar tu imagen personalizada con plugins: redis, cloudflare, ratelimit
    image: registry.gitlab.com/mls_dev/caddy_proxy_docker_cloudflare:latest 
    ports:
      # Cambiar la sintaxis corta por la larga con 'mode: host'
      - target: 80
        published: 80
        protocol: tcp
        mode: host
      - target: 443
        published: 443
        protocol: tcp
        mode: host
      - target: 443
        published: 443
        protocol: udp
        mode: host
    environment:
      - CADDY_INGRESS_NETWORKS=network_general_internal
      - CLOUDFLARE_API_TOKEN={env.CLOUDFLARE_API_TOKEN}
      - ACME_EMAIL={env.emailcloudflare}
    networks:
      - network_general_internal
      - monitor_monitoring
    volumes:
      - /var/run/docker.sock:/var/run/docker.sock
    stop_grace_period: 5s      
    deploy:
      mode: global

      update_config:
        parallelism: 1
        order: stop-first
        failure_action: rollback
        delay: 10s
      rollback_config:
        parallelism: 1
        order: stop-first      
      labels:

        caddy_0.storage: "redis"
        caddy_0.storage.address: "redis_caddy:6379"
        caddy_0.storage.key_prefix: "caddytls"        
        caddy_0.storage.timeout: "10"        
        caddy_0.acme_dns: "cloudflare {env.CLOUDFLARE_API_TOKEN}"
        caddy_0.order: "rate_limit before reverse_proxy"
        caddy_global.rate_limit: "{remote.host}"
        caddy_global.rate_limit.events: "30"
        caddy_global.rate_limit.window: "1s"
        caddy_global.tls.dns: "cloudflare {env.CLOUDFLARE_API_TOKEN}"

#        caddy_0.log: ""
#        caddy_0.log.output: "stdout"
#        caddy_0.log.level: "DEBUG"

#        com.github.mazzolino.shepherd.enable: "true" # ahora redesplegar caddy automaticamenete pierde el servicio
    healthcheck:
      test: wget --no-verbose --tries=1 --spider http://127.0.0.1:2019/metrics || exit 1
      interval: 10s
      timeout: 1s
      retries: 3
      
volumes:
  redis_data_caddy:

networks:
  network_general_internal:
    external: true
    name: network_general_internal
  monitor_monitoring:
    external: true
    name: monitor_monitoring