fix(emqx_mgmt): format stoped nodes correctly

[zmstone]

fixes: #10034

overview:

node view:

sample output

[
  {
    "connections": 0,
    "edition": "Opensource",
    "load1": 0.6,
    "load15": 0.89,
    "load5": 1.01,
    "log_path": "log.file_handler.default.enable is false, not logging to file",
    "max_fds": 1048576,
    "memory_total": "62.75G",
    "memory_used": "10.92G",
    "node": "emqx@node2.emqx.io",
    "node_status": "running",
    "otp_release": "24.3.4.2-2/12.3.2.2",
    "process_available": 2097152,
    "process_used": 625,
    "role": "core",
    "sys_path": "/opt/emqx",
    "uptime": 97457,
    "version": "5.0.18-ge8427c99"
  },
  {
    "node": "emqx@node1.emqx.io",
    "node_status": "stopped"
  }
]

[qzhuyan]

This assumes we only have one logger handler that has file backend.

[zmstone]

it does not assume that, it returns the first one.