ECONNREFUSED stops any attempted reconnections on initial connect

[delize]

• Node version: Unknown
• Browser version: Version 122.0.6261.69
• Device, operating system: Docker Image
• The Lounge version: v4.4.1

---

I am running:

• ZNC
• TL

In a docker compose file:

---
version: '3.8'
services:
  thelounge:
    image: thelounge/thelounge:latest
    container_name: thelounge
    user: 1001:1001
    networks:
      - irc
      - backend
    # ports:
    #   - '9001:9000'
    restart: unless-stopped
    volumes:
      - ${BASE_PATH}/thelounge:/var/opt/thelounge
    depends_on:
      - znc
    environment:
      - TZ=${TIME_ZONE:-Europe/Stockholm}
  znc:
    image: znc:latest
    container_name: znc
    user: ${PUID}
    networks:
      - irc
      - backend
    environment:
      - TZ=${TIME_ZONE:-Europe/Stockholm}
      - uid=${PGID}
      - gid=${PGID}
    # ports:
      # - '7000:7000'
      # - '65500:65500'
    volumes:
      - ${BASE_PATH}/znc:/znc-data
      - ${BASE_PATH}/swag/etc/letsencrypt/live/${DOMAIN}/:/le-ssl:ro
    restart: unless-stopped
networks:
  irc:
    driver: bridge
    name: irc
  backend:
    external: true

When I have TL connect to the ZNC server, the first network in TL always fails to connect with:

21:55:02 Network created, connecting to znc:6667...
21:55:02 *** Connection closed unexpectedly: Error: connect ECONNREFUSED 172.19.0.11:6667
21:55:02 Disconnected from the network, and will not reconnect. Use /connect to reconnect again.

After this point, the network never attempts to reconnect.

Where as the second Network connects within <1 second:

21:55:02 Network created, connecting to znc:6667...
21:55:03 Connected to the network.
21:55:03 Enabled capabilities: batch, cap-notify, echo-message, multi-prefix, server-time, userhost-in-names, [znc.in/self-message](http://znc.in/self-message), [znc.in/server-time-iso](http://znc.in/server-time-iso)

Now I can / and need to - write a healthcheck into znc / thelounge based on this: https://docs.docker.com/compose/compose-file/05-services/#long-syntax-1

But, I do think that TL should attempt to reconnect after a period of time on an initial connection attempt, similar to what is occurring here: #4159