服务端消费者,偶现重连

服务端的消息者,在高峰期间,每隔几分钟,会出现几次,报错说是 timeout, 这个大概是什么原因造成的,有什么排查和解决问题的方向吗?

150 左右生产者在线
消息发送和接受频率 100条/秒

2025-10-23T09:35:29.643652+08:00 [warning] clientid: center_api_server, msg: socket_error, peername: 172.19.28.191:39566, username: center_api_server, reason: timeout
2025-10-23T09:35:29.644272+08:00 [warning] clientid: center_api_server, msg: socket_error, peername: 172.19.28.191:39566, username: center_api_server, reason: enotconn
2025-10-23T09:35:29.644658+08:00 [warning] clientid: center_api_server, msg: socket_error, peername: 172.19.28.191:39566, username: center_api_server, reason: enotconn
2025-10-23T09:35:29.644950+08:00 [warning] clientid: center_api_server, msg: socket_error, peername: 172.19.28.191:39566, username: center_api_server, reason: enotconn
2025-10-23T09:35:29.645072+08:00 [warning] clientid: center_api_server, msg: socket_error, peername: 172.19.28.191:39566, username: center_api_server, reason: enotconn
2025-10-23T09:35:29.645542+08:00 [warning] clientid: center_api_server, msg: socket_error, peername: 172.19.28.191:39566, username: center_api_server, reason: enotconn
2025-10-23T09:35:29.645706+08:00 [warning] clientid: center_api_server, msg: socket_error, peername: 172.19.28.191:39566, username: center_api_server, reason: enotconn
2025-10-23T09:35:29.645842+08:00 [warning] clientid: center_api_server, msg: socket_error, peername: 172.19.28.191:39566, username: center_api_server, reason: enotconn
2025-10-23T09:35:29.646111+08:00 [warning] clientid: center_api_server, msg: socket_error, peername: 172.19.28.191:39566, username: center_api_server, reason: enotconn
2025-10-23T09:35:29.646279+08:00 [warning] clientid: center_api_server, msg: socket_error, peername: 172.19.28.191:39566, username: center_api_server, reason: enotconn
2025-10-23T09:35:29.646383+08:00 [warning] clientid: center_api_server, msg: socket_error, peername: 172.19.28.191:39566, username: center_api_server, reason: enotconn
2025-10-23T09:35:29.646529+08:00 [warning] clientid: center_api_server, msg: socket_error, peername: 172.19.28.191:39566, username: center_api_server, reason: enotconn
2025-10-23T09:35:29.646656+08:00 [warning] clientid: center_api_server, msg: socket_error, peername: 172.19.28.191:39566, username: center_api_server, reason: enotconn
2025-10-23T09:35:29.646778+08:00 [warning] clientid: center_api_server, msg: socket_error, peername: 172.19.28.191:39566, username: center_api_server, reason: enotconn
2025-10-23T09:35:29.646912+08:00 [warning] clientid: center_api_server, msg: socket_error, peername: 172.19.28.191:39566, username: center_api_server, reason: enotconn
2025-10-23T09:35:29.647018+08:00 [warning] clientid: center_api_server, msg: socket_error, peername: 172.19.28.191:39566, username: center_api_server, reason: enotconn
  1. 首先出现 timeout(第一条)- 连接超时
  2. 随后连续出现多个 enotconn - 在已断开的连接上尝试操作

原因

这是一个典型的 客户端连接断开但未正确处理 的场景:

  1. 初始问题:客户端 center_api_server (IP: 172.19.28.191:39566) 的连接超时
  • 可能是网络延迟、丢包
  • 可能是 MQTT keepalive 超时
  • 可能是服务端或客户端负载过高

多搞点客户端来共享消费。

业务低峰时期,客户端 center_api_server 一直正常,无 timeout
业务高峰时期,才会出现,有可能是接受消息的方法 messageArrived ,处理积压导致的吗,用的是 org.eclipse.paho.mqttv5.client; 客户端

paho.mqttv5.client 客户端,心跳线程与业务处理线程是共用的同一个线程吗

可以去sdk的社区问一下,我不懂java

恩恩,感谢:pray: