[rabbitmq-discuss] Publisher throttled way too early

[François Beausoleil]

Hello Simon!

Le lundi 16 avril 2012 à 06:36, Simon MacMullen a écrit :

>  
> So therefore I think that your system can really only handle that many  
> messages per second. Are they persistent, or are your queues big enough  
> that non-persistent messages page out? That sort of performance sounds  
> more like the sort of thing you would see when IO-bound.
>  

I'm a bit chicken, and I use durable queues and persistent messages, because I'm on a VM. With your other clarifications, and the fact I can barely sustain 300 transactions per second on a similar host on another VM using PostgreSQL, I suspect you are correct.

Since this is reloading historical data for reprocessing, I'll just skip the persistence for only those messages.
  
Thanks!
--
François Beausoleil