[rabbitmq-discuss] RabbitMQ: Different consumer technique
mcintoshj at gmail.com
Mon Mar 10 16:15:44 GMT 2014
Seems like a lot of heavy lifting for a consumer process. If you use the
alternative where rabbit pushes changes to your consumer (instead of
basic_get), then rabbit will push new messages as they come in to your
consumer, and the consumer will be idle otherwise using minimal resources.
For "basic.get_batch" - that's exactly what the prefetch count does on
consumers - get 10 messages at once and process them. Is there some
particularly need to wrap it in all the extra layers?
On Mon, Mar 10, 2014 at 10:43 AM, Ajinkya Apte <aapte at zynga.com> wrote:
> (First time user of RabbitMQ)
> Rather than writing a standalone PHP consumer script that keeps consuming
> the queue, I was planning on writing a batch get and quit PHP script. All
> that script does is get a batch of 10 from the queue, if the batch isn't
> available the script immediately exits. If available, it processes the
> batch and quits. I use a python script to make multiple calls to my 'Get
> Batch script' and make use of apache workers to do the work rather than
> making the php standalone script do the work.
> 1. Firstly, do you see a reason why I should not think about consuming the
> queue this way?
> 2. (I am using the amq php lib) How can I do a batch get and exit if the
> batch is not present. With what I have learnt seems like I can set the qos
> to my batch count and use batch_consume for this. But if the batch is not
> available the script seems to wait till the batch is available. Is there a
> way I can timeout?
> 3. Do you have any usage recommendation for me?
> 4. Any plans of having basic.get_batch?
> Sent from my iPhone
> rabbitmq-discuss mailing list
> rabbitmq-discuss at lists.rabbitmq.com
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the rabbitmq-discuss