RabbitMQ与.net core(四) 消息的优先级 与 死信队列

1.消息的优先级

假如现在有个需求，我们需要让一些优先级最高的通知推送到客户端，我们可以使用redis的sortedset，也可以使用我们今天要说的rabbit的消息优先级属性

Producer代码

using RabbitMQ.Client;using System;using System.Collections.Generic;using System.Text;using System.Threading;using System.Threading.Tasks;namespace RabbitMQConsole
{    class Program
    {        static void Main(string[] args)
        {
            ConnectionFactory factory = new ConnectionFactory();
            factory.HostName = "39.**.**.**";
            factory.Port = 5672;
            factory.VirtualHost = "/";
            factory.UserName = "root";
            factory.Password = "root";            var exchange = "change4";            var route = "route2";            var queue9 = "queue9";            using (var connection = factory.CreateConnection())
            {                using (var channel = connection.CreateModel())
                {
                    channel.ExchangeDeclare(exchange, type: "fanout", durable: true, autoDelete: false);
　　　　　　　　　　　　//x-max-priority属性必须设置，否则消息优先级不生效
                    channel.QueueDeclare(queue9, durable: true, exclusive: false, autoDelete: false,arguments: new Dictionary<string, object> { { "x-max-priority", 50 } });
                    channel.QueueBind(queue9, exchange, queue9);                    while(true)
                    {                        var messagestr = Console.ReadLine();                        var messagepri = Console.ReadLine();                        var props = channel.CreateBasicProperties();
                        props.Persistent = true;
                        props.Priority = (byte)int.Parse(messagepri);//设置消息优先级
                        channel.BasicPublish(exchange, route, true, props, Encoding.UTF8.GetBytes(messagestr));
                    }
                }
            }
        }
    }
}

consumer代码

using RabbitMQ.Client;using RabbitMQ.Client.Events;using System;using System.Collections.Generic;using System.Text;using System.Threading;namespace RabbitMQClient
{    class Program
    {        private static readonly ConnectionFactory rabbitMqFactory = new ConnectionFactory()
        {
            HostName = "39.**.**.**",
            Port = 5672,
            UserName = "root",
            Password = "root",
            VirtualHost = "/"
        };        static void Main(string[] args)
        {            var exchange = "change4";            var route = "route2";            var queue9 = "queue9";            using (IConnection conn = rabbitMqFactory.CreateConnection())            using (IModel channel = conn.CreateModel())
            {
                channel.ExchangeDeclare(exchange, "fanout", durable: true, autoDelete: false);
                channel.QueueDeclare(queue9, durable: true, exclusive: false, autoDelete: false, arguments: new Dictionary<string, object> { { "x-max-priority", 50 } });
                channel.QueueBind(queue9, exchange, route);

                channel.BasicQos(prefetchSize: 0, prefetchCount: 50, global: false);
                EventingBasicConsumer consumer = new EventingBasicConsumer(channel);
                consumer.Received += (model, ea) =>
                {
                    Byte[] body = ea.Body;
                    String message = Encoding.UTF8.GetString(body);
                    Console.WriteLine( message);
                    channel.BasicAck(deliveryTag: ea.DeliveryTag, multiple: false);
                };

                channel.BasicConsume(queue: queue9, autoAck: false, consumer: consumer);
                Console.ReadLine();
            }
        }
    }
}

运行producer

在运行consumer

可以看出消息是按优先级消费的

回到顶部

2.死信队列

死信队列可以用来做容错机制，当我们的消息处理异常时我们可以把消息放入到死信队列中，以便后期处理，死信的产生有三种

1.消息被拒(basic.reject or basic.nack)并且没有重新入队(requeue=false)；

2.当前队列中的消息数量已经超过最大长度。

3.消息在队列中过期，即当前消息在队列中的存活时间已经超过了预先设置的TTL(Time To Live)时间；

看代码

using RabbitMQ.Client;using System;using System.Collections.Generic;using System.Text;using System.Threading;using System.Threading.Tasks;namespace RabbitMQConsole
{    class Program
    {        static void Main(string[] args)
        {
            ConnectionFactory factory = new ConnectionFactory();
            factory.HostName = "39.**.**.**";
            factory.Port = 5672;
            factory.VirtualHost = "/";
            factory.UserName = "root";
            factory.Password = "root";            var exchangeA = "changeA";            var routeA = "routeA";            var queueA = "queueA";            var exchangeD = "changeD";            var routeD = "routeD";            var queueD = "queueD";            using (var connection = factory.CreateConnection())
            {                using (var channel = connection.CreateModel())
                {
                    channel.ExchangeDeclare(exchangeD, type: "fanout", durable: true, autoDelete: false);
                    channel.QueueDeclare(queueD, durable: true, exclusive: false, autoDelete: false);
                    channel.QueueBind(queueD, exchangeD, routeD);

                    channel.ExchangeDeclare(exchangeA, type: "fanout", durable: true, autoDelete: false);
                    channel.QueueDeclare(queueA, durable: true, exclusive: false, autoDelete: false, arguments: new Dictionary<string, object> {
                                         { "x-dead-letter-exchange",exchangeD}, //设置当前队列的DLX
                                         { "x-dead-letter-routing-key",routeD}, //设置DLX的路由key，DLX会根据该值去找到死信消息存放的队列
                                         { "x-message-ttl",10000} //设置消息的存活时间，即过期时间                                         });
                    channel.QueueBind(queueA, exchangeA, routeA);                    var properties = channel.CreateBasicProperties();
                    properties.Persistent = true;                    //发布消息                    channel.BasicPublish(exchange: exchangeA,
                                         routingKey: routeA,
                                         basicProperties: properties,
                                         body: Encoding.UTF8.GetBytes("message"));
                }
            }
        }
    }
}

这样10秒后消息过期，我们可以看到queueD中有了消息

原文出处：https://www.cnblogs.com/chenyishi/p/10242162.html