MCP的SSE重连机制，低成本接入框架，快速上手SSE 重连机制定时任务线程池：定期 ping MCP Server

SSE 重连机制

[!TIP] MCP 的 SSE 连接，Client 侧和 Server 侧一段时间未传输数据会断开，或者 Server 侧挂掉重启后，Client 侧此时无法建立 SSE 连接。这里提供了一个自动重连机制，开箱即用，基本原理如下：定义两个线程池

定时任务线程池：定期 ping MCP Server 侧，若请求失败则记录至延时队列
异步线程池：从延时队列中拉取失败的 MCP Server 侧，建立连接

框架代码已提供至社区：github.com/alibaba/spr…

快速上手例子：github.com/GTyingzi/sp… 下的 mcp/client 目录下的 mcp-recovery-client 模块

快速上手

pom.xml

<properties>
    <!-- Spring AI Alibaba -->
    <spring-ai-alibaba.version>1.0.0.3-SNAPSHOT</spring-ai-alibaba.version>
</properties>

<dependencies>

    <dependency>
        <groupId>org.springframework.ai</groupId>
        <artifactId>spring-ai-autoconfigure-model-openai</artifactId>
    </dependency>

    <dependency>
        <groupId>org.springframework.ai</groupId>
        <artifactId>spring-ai-autoconfigure-model-chat-client</artifactId>
    </dependency>

    <dependency>
        <groupId>org.springframework.boot</groupId>
        <artifactId>spring-boot-starter-web</artifactId>
    </dependency>

    <dependency>
        <groupId>org.springframework.ai</groupId>
        <artifactId>spring-ai-starter-mcp-client-webflux</artifactId>
    </dependency>

    <dependency>
        <groupId>com.alibaba.cloud.ai</groupId>
        <artifactId>spring-ai-alibaba-starter-mcp-recovery-client</artifactId>
        <version>${spring-ai-alibaba.version}</version>
    </dependency>

</dependencies>

application.yml

server:
  port: 19100

spring:
  application:
    name: mcp-recovery-client
  ai:
    openai:
      api-key: ${DASHSCOPEAPIKEY}
      base-url: https://dashscope.aliyuncs.com/compatible-mode
      chat:
        options:
          model: qwen-max
    mcp:
      client:
        enabled: false
        name: my-mcp-client
        version: 1.0.0
        request-timeout: 600s
        type: ASYNC  # or ASYNC for reactive applications
        sse:
          connections:
            server1:
              url: http://localhost:19000 # 本地

    alibaba:
      mcp:
        recovery:
          enabled: true
          delay: 5s
          stop: 10s

# 打印日志
logging:
  level:
    com:
      alibaba:
        cloud:
          ai:
            mcp:
              client: DEBUG

spring.ai.mcp.client.enabled: false：设为 false，取消 McpSyncClient 或 McpAsyncClient 配置
spring.ai.alibaba.mcp.recovery：重连配置
- enabled：true 为启动，默认为 false
- delay：MCP Client 侧多久 Ping 一次 MCP Server，默认为 5 秒
- stop：MCP 关闭线程池的最大等待时间，默认为 10 秒

controller

package com.spring.ai.tutorial.mcp.client.controller;

import com.alibaba.cloud.ai.mcp.client.McpAsyncRecovery;
import org.springframework.ai.chat.client.ChatClient;
import org.springframework.web.bind.annotation.GetMapping;
import org.springframework.web.bind.annotation.RequestMapping;
import org.springframework.web.bind.annotation.RequestParam;
import org.springframework.web.bind.annotation.RestController;

/**
 * @author yingzi
 * @since 2025/7/15
 */
@RestController
@RequestMapping("/time")
public class TimeController {

    private final ChatClient chatClient;

    private final McpAsyncRecovery mcpAsyncRecovery;

    public TimeController(ChatClient.Builder chatClientBuilder, McpAsyncRecovery mcpAsyncRecovery) {
        this.chatClient = chatClientBuilder.build();
        this.mcpAsyncRecovery = mcpAsyncRecovery;
    }

    @GetMapping("/chat")
    public String chatTime(@RequestParam(value = "query", defaultValue = "请告诉我现在北京时间几点了") String query) {
        return chatClient.prompt(query)
                .toolCallbacks(mcpAsyncRecovery.getToolCallback())
                .call()
                .content();
    }
}