Runner with only IPv6

When setting public_ipv4_disabled = true on the plugin configuration, it seems the coordinator/manager cannot connect to the newly created runner. The runner is created without any problems, but it then hangs at Dialing instance 1234564789....

The manger has both an IPv4 and an IPv6 configured. The manager has two firewall rules for only allowing incoming traffic on port 22 and for ICMP requests.

This might be mitigated with the private_networks property, but not testet.

Configuration:

[[runners]]
  name = "hetzner-ci-autoscaler-0"
  url = "https://gitlab.com"
  id = 123456789
  token = ""
  token_obtained_at = 2024-10-17T08:58:43Z
  token_expires_at = 0001-01-01T00:00:00Z
  executor = "docker-autoscaler"

  output_limit = 12288
  request_concurrency = 5

  [runners.docker]
    tls_verify = false
    image = "alpine:latest"
    volumes = ["/cache"]

    privileged = true
    disable_entrypoint_overwrite = false
    oom_kill_disable = false
    disable_cache = false
    shm_size = 0
    network_mtu = 0

  [runners.autoscaler]
    capacity_per_instance = 1
    max_use_count = 0
    max_instances = 5
    plugin = "hetznercloud/fleeting-plugin-hetzner:latest"
    update_interval = "1m0s"
    update_interval_when_expecting = "5s"

    [runners.autoscaler.plugin_config]
      image = "docker-ce"
      location = "fsn1"
      name = "runner-hetzner-ci"
      public_ipv4_disabled = true
      private_networks = []
      server_type = "cpx41"
      token = ""

    [runners.autoscaler.scale_throttle]
      limit = 10

    [runners.autoscaler.connector_config]
      use_external_addr = true
      os = "linux"
      protocol = "ssh"