Article directory
foreword
I believe that many students have developed WEB services. In the development of WEB services, caching, queues, read-write separation, peak-shaving and valley-filling, current limiting and downgrading are generally used to improve service performance and ensure normal use of services. For peak shaving and valley filling, we have to use our MQ message middleware, such as kafka suitable for big data, rabbitmq with high performance and high transaction activity, etc. The selection and integration of MQ is already indispensable in JAVA WEB development missing part. Of course, Spring Cloud, known as the family bucket of the JAVA microservice framework, also provides a good function of adapting middleware. Today we will integrate the message-driven Spring Cloud Stream provided by Spring Cloud, a family bucket of microservices.
A Brief Analysis of Spring Cloud Stream
Spring Cloud Stream is a framework for building microservices with message-driven capabilities. Applications interact with binders through inputs and outputs channels, and binders communicate with message middleware.
The role of the binder is to glue the message middleware, which is equivalent to encapsulating and integrating the third-party middleware, so that developers do not need to care about how the underlying message middleware operates.
inputs is the message input channel, similar to the consumer of the message middleware; outputs is the message output channel, similar to the producer of the message middleware. The application sends and receives messages no longer directly calls the interface or logic code of the message middleware, but directly uses the OUTPUT and INPUT channels of Spring Cloud Stream for processing.
Various message middleware can be selected through binder binding, and the relevant parameters of the middleware can be configured through binding, so that the application can flexibly configure and switch message middleware.
Integration of Spring Cloud Stream and rabbitmq
This integration is directly integrated with rabbitmq. If you are using kafka, you can directly transplant the configuration and modify the corresponding bonding mq.
This integration has added a consumption retry mechanism, a dead letter queue, and provided a dead letter queue consumption monitoring method, which can be directly transplanted to the production environment.
1. Add pom dependencies
The introduction of spring-cloud-starter-stream-rabbit needs to be introduced from Spring Cloud, pay attention to the configuration of dependencyManagement.
<properties>
<java.version>1.8</java.version>
<spring-cloud.version>Hoxton.SR10</spring-cloud.version>
</properties>
<dependencies>
<dependency>
<groupId>org.springframework.cloud</groupId>
<artifactId>spring-cloud-starter-stream-rabbit</artifactId>
</dependency>
</dependencies>
<dependencyManagement>
<dependencies>
<dependency>
<groupId>org.springframework.cloud</groupId>
<artifactId>spring-cloud-dependencies</artifactId>
<version>${spring-cloud.version}</version>
<type>pom</type>
<scope>import</scope>
</dependency>
</dependencies>
</dependencyManagement>
2. Application.yml adds mq configuration
spring:
rabbitmq:
host: 127.0.0.1
port: 5672
username: admin
password: admin
virtual-host: /
cloud:
stream:
binders: #stream框架粘接的mq
myRabbit: #自定义个人mq名称
type: rabbit
environment:
spring:
rabbitmq:
host: 127.0.0.1
port: 5672
username: admin
password: admin
virtual-host: /
bindings: #stream绑定信道
output_channel: #自定义发送信道名称
destination: assExchange #目的地 交换机/主题
content-type: application/json
binder: myRabbit #粘接到的mq
group: assGroup
input_channel: #自定义接收信道
destination: assExchange #目的地 交换机/主题
content-type: application/json
binder: myRabbit #粘接到的mq
group: assGroup
consumer:
maxAttempts: 3 # 尝试消费该消息的最大次数（消息消费失败后，发布者会重新投递）。默认3
backOffInitialInterval: 1000 # 重试消费消息的初始化间隔时间。默认1s，即第一次重试消费会在1s后进行
backOffMultiplier: 2 # 相邻两次重试之间的间隔时间的倍数。默认2
backOffMaxInterval: 10000 # 下一次尝试重试的最大时间间隔，默认为10000ms，即10s
rabbit: #stream mq配置
bindings:
input_channel:
consumer:
concurrency: 1 #消费者数量
max-concurrency: 5 #最大消费者数量
durable-subscription: true #持久化队列
recovery-interval: 3000 #3s 重连
acknowledge-mode: MANUAL #手动
requeue-rejected: false #是否重新放入队列
auto-bind-dlq: true #开启死信队列
requeueRejected: true #异常放入死信
3. Define the input and output channels
/**
* MqChannel
* @author senfel
* @version 1.0
* @date 2023/6/2 15:46
*/
public interface MqChannel {
/**
* 消息目的地 RabbitMQ中为交换机名称
*/
String destination = "assExchange";
/**
* 输出信道
*/
String OUTPUT_CHANNEL = "output_channel";
/**
* 输入信道
*/
String INPUT_CHANNEL = "input_channel";
/**
* 死信队列
*/
String INPUT_CHANNEL_DLQ = "assExchange.assGroup.dlq";
@Output(MqChannel.OUTPUT_CHANNEL)
MessageChannel output();
@Input(MqChannel.INPUT_CHANNEL)
SubscribableChannel input();
}
4. Use input and output channels to send and receive messages
TestMQService
/**
* TestMQService
* @author senfel
* @version 1.0
* @date 2023/6/2 15:47
*/
public interface TestMQService {
/**
* 发送消息
*/
void send(String str);
}
TestMQServiceImpl
/**
* TestMQServiceImpl
* @author senfel
* @version 1.0
* @date 2023/6/2 15:49
*/
@Service
@Slf4j
@EnableBinding(MqChannel.class)
public class TestMQServiceImpl implements TestMQService {
@Resource
private MqChannel mqChannel;
@Override
public void send(String str) {
mqChannel.output().send(MessageBuilder.withPayload("测试=========="+str).build());
}
/**
* 接收消息监听
* @param message 消息体
* @param channel 信道
* @param tag 标签
* @param death
* @author senfel
* @date 2023/6/5 9:25
* @return void
*/
@StreamListener(MqChannel.INPUT_CHANNEL)
public void process(String message,
@Header(AmqpHeaders.CHANNEL) Channel channel,
@Header(AmqpHeaders.DELIVERY_TAG) long tag) throws Exception {
log.info("message : "+message);
if(message.contains("9")){
// 参数1为消息的tag 参数2为是否多条处理 参数3为是否重发
//channel.basicNack(tag,false,false);
System.err.println("--------------消费者消费异常--------------------------------------");
System.err.println(message);
throw new RuntimeException("抛出异常");
}else{
System.err.println("--------------消费者--------------------------------------");
System.err.println(message);
channel.basicAck(tag,false);
}
}
/**
* 死信监听
* @param message 消息体
* @param channel 信道
* @param tag 标签
* @param death
* @author senfel
* @date 2023/6/5 14:30
* @return void
*/
@RabbitListener(
bindings = @QueueBinding(
value = @Queue(MqChannel.INPUT_CHANNEL_DLQ)
, exchange = @Exchange(MqChannel.destination)
),
concurrency = "1-5"
)
public void processByDlq(String message,
@Header(AmqpHeaders.CHANNEL) Channel channel,
@Header(AmqpHeaders.DELIVERY_TAG) long tag) throws Exception {
log.info("message : "+message);
System.err.println("---------------死信消费者------------------------------------");
System.err.println(message);
}
}
controller
/**
* @author senfel
* @version 1.0
* @date 2023/6/2 17:27
*/
@RestController
public class TestController{
@Resource
private TestMQService testMQService;
@GetMapping("/test")
public String testMq(String str){
testMQService.send(str);
return str;
}
}
5. Simulate normal message consumption
6. Simulate exception message
The abnormal message enters dead letter consumption after retrying to meet 3 deliveries
