Golang高吞吐量RabbitMQ消费：如何高效批量获取消息并统一确认？

// ... RabbitMQ连接和channel初始化 ...

ch.Qos(prefetchCount, 0, false) // 设置预取数量

messages := make([]amqp.Delivery, 0, bufferSize) // 消息缓冲区
timer := time.NewTimer(1 * time.Second) // 定时器，辅助批量处理

for {
    select {
    case d := <-ch.Consume(...): // 接收消息
        messages = append(messages, d)
        if len(messages) >= bufferSize {
            processAndAck(messages)
            messages = messages[:0] // 清空缓冲区
        }
    case <-timer.C:
        if len(messages) > 0 {
            processAndAck(messages)
            messages = messages[:0]
        }
        timer.Reset(1 * time.Second) // 重置定时器
    }
}

func processAndAck(deliveries []amqp.Delivery) {
    // 批量处理消息
    successfulDeliveries := make([]amqp.Delivery, 0, len(deliveries))
    failedDeliveries := make([]amqp.Delivery, 0, len(deliveries))
    for _, delivery := range deliveries {
        if err := processMessage(delivery); err != nil {
            failedDeliveries = append(failedDeliveries, delivery)
        } else {
            successfulDeliveries = append(successfulDeliveries, delivery)
        }
    }

    // 统一确认
    if len(successfulDeliveries) > 0 {
        ch.AckMultiple(successfulDeliveries[len(successfulDeliveries)-1].DeliveryTag, false)
    }
    if len(failedDeliveries) > 0 {
        ch.NackMultiple(failedDeliveries[len(failedDeliveries)-1].DeliveryTag, false, true) // requeue: true
    }
}

func processMessage(delivery amqp.Delivery) error {
    // 处理单个消息逻辑
    // ...
    return nil // 或返回错误
}

// ... 错误处理，重试机制，死信队列等 ...