The broker quit operation - "Keep alive time overdue".

The broker quit operation after a client disconnects. It happened a few times, but it is not something I can reproduce. Below is all I could catch.

Using nmqtt-1.0.4, Nim 1.6.6, Windows 11

rx> Connect(00): 00 04 4D 51 54 54 04 06 00 3C 00 19 73 68 65 6C 6C 79 70 6C 75 67 2D 73 2D 42 43 46 46 34 44 35 41 32 36 30 37 00 29 73 68 65 6C 6C 69 65 73 2F 73 68 65 6C 6C 79 70 6C 75 67 2D 73 2D 42 43 46 46 34 44 35 41 32 36 30 37 2F 6F 6E 6C 69 6E 65 00 05 66 61 6C 73 65
tx> ConnAck(00): 00 02
Connections >> shellyplug-s-BCFF4D5A2607 was disconnected. Keep alive time overdue.
Client      >> shellyplug-s-BCFF4D5A2607
  host:
  sslOn: false
  sslCert:
  sslKey:
  verbosity: 0
  beenConnected: true
  username:
  password:
  clientId: shellyplug-s-BCFF4D5A2607
  s: ..disconnected..
  msgIdSeq: 3
  workQueue: 0
  pubCallbacks: 0
  inWork: false
  keepAlive: 60
  willFlag: false
  willQoS: 0
  willRetain: false
  willTopic: shellies/shellyplug-s-BCFF4D5A2607/online
  willMsg: false
  proto: MQTT
  version: 4
  connFlags: 00000110
  subscribed: 3

PS C:\Users\xxx>```

No source file, test, or entry point is named. Start by tracing the broker's keep-alive timeout handling and compare it with the captured Connect and ConnAck exchange and the reported client disconnect. Done means identifying the cause, preventing the broker from quitting incorrectly, and demonstrating the behavior with a reproducible regression test.