On Thu, Aug 20, 2009 at 2:35 AM, voomstudio.com<[email protected]> wrote:
> Will be good to see how performance holds up
> - are you able to give me any benchmarks as to what kind of loads
> beanstalkd can handle. Eg I thinking of around 5000 jobs a second with
> a small json payload.

Throughput depends strongly on the hardware and network. 5000 jobs per
second is typical.

The first bottleneck is usually the workers. Even a worker that just
reserves and deletes jobs (with no other processing) is usually the
slowest part.

If you have many workers on many machines, the next bottleneck is
usually the network. Beanstalkd can often saturate a network link
without chewing up much cpu.

> Curious as to if I will have to create multiple
> tubes to deal with load - or if this is a non issue. Any reading is
> appreciated.

It's a non-issue. The performance of beanstalkd with several tubes
won't be noticeably different from a single tube. (However, the entire
*system* is likely to be more efficient with a single tube, since your
workers won't waste time waiting on empty tubes while jobs are ready
in other tubes.)

> Thanks for all your help Keith - first thoughts are great & your
> responses have been accurate & timely.

No problem. Thanks.

kr

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups 
"beanstalk-talk" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to 
[email protected]
For more options, visit this group at 
http://groups.google.com/group/beanstalk-talk?hl=en
-~----------~----~----~----~------~----~------~--~---

Reply via email to