ActiveMQ 的 prefetch-limit

ActiveMQ 的 prefetch-limit

 

1.官方文档

http://activemq.apache.org/what-is-the-prefetch-limit-for.html

 

One of the aims of ActiveMQ is to be a high performance message bus. This  means

using a  SEDA(http://activemq.apache.org/seda.html) architecture  to perform  as

much work as possible asynchronously. To be  able to  achieve high   performance

it  is important   to stream  messages  to consumers  as  fast  as possible   so

that   the consumer   always has   a buffer   of messages,  in  RAM,  ready  to

process  -  rather  than  have  them  explicitly  pull  messages from the server

which adds significant latency per message.

 

ActiveMQ的一个目标是成为高性能消息总线.这就是说,ActiveMQ采用SEDA结构以异步方式尽可能

高效的处理更多工作.为了获取高性能,ActiveMQ必须尽可能快的将消息传送给消费者以便消费者

可以使用RAM消息缓存,而不是让消费者主动从代理服务器一个一个地拉取消息,因为这样会为每个

消息增加大量的延迟.

 

There  is a  danger however  that this  aggressive pushing  of messages  to the

consumers  could  flood a  consumer  as typically  its  much faster  to  deliver

messages to the consumer than it often is to actually process them.

 

这种侵入式的推送消息给消费者的行为是危险的,因为可能导致消费者被大量消息淹没.毕竟,通常

发送消息给消费者要比消费者处理消息快的多.

 

So ActiveMQ  uses a  prefetch limit  on how  many messages  can be streamed to a

consumer at  any point  in time.  Once the  prefetch limit  is reached,  no more

messages are dispatched to the  consumer until the consumer starts  sending back

acknowledgements of messages (to indicate that the message has been  processed).

The actual prefetch limit value can be specified on a per consumer basis.

 

因此,ActiveMQ使用prefetch limit限制任何时候消息消费者已接受到但尚未完成处理的消息数量.

一旦消息数量达到prefetch limit限制,则在消费者开始发送消息确认(指示消息已经被处理了)之前

不会再发送任何消息给消费者.可以为每一个消息消费者单独设置prefetch limit 值.

 

Its a good  idea to have  large values of  the prefetch limit  if you want  high

performance and if you have high message volumes. If you have very few  messages

and each message  takes a very  long time to  process you might  want to set the

prefetch  value  to 1  so  that a  consumer  is given  one  message at  a  time.

Specifying  a prefetch  limit of  zero means  the consumer  will poll  for more

messages, one at a time, instead of the message being pushed to the consumer.

 

如果你打算高效的发送大量消息,那么为prefetch limit设置一个比较大的值是一个不错的注意.

如果你只打算发送很少的消息,并且每一个消息的处理时间很长,那么可以设置prefetch limit值

为1,这样消息者每次只会接收到一个消息.设置prefetch  limit 值为0表示,消费者每次主动拉取更

多消息,而不是由代理推送消息到消费者.

 

Specifying the PrefetchPolicy

配置 prefetch  limit

 

You   can   specify   an   instance   of   the   ActiveMQPrefetchPolicy   on  an

ActiveMQConnectionFactory or  ActiveMQConnection. This  allows you  to configure

all the individual prefetch values; as  each different quality of service has  a

different value. e.g.

 

persistent queues (default value: 1000)

non-persistent queues (default value: 1000)

persistent topics (default value: 100)

non-persistent topics (default value: Short.MAX_VALUE -1)

 

It can also be configured on the connection URI used when establishing a connection the broker:

To change the prefetch size for all consumer types you would use a connection URI similar to:

tcp://localhost:61616?jms.prefetchPolicy.all=50

 

To change the prefetch size for just queue consumer types you would use a connection URI similar to:

tcp://localhost:61616?jms.prefetchPolicy.queuePrefetch=1

 

It can also be configured on a per consumer basis using Destination Options.

queue = new ActiveMQQueue("TEST.QUEUE?consumer.prefetchSize=10");

consumer = session.createConsumer(queue);