On Sunday, October 20, 2013 11:14:40 PM UTC+1, Oto Iashvili wrote:
I would like to understand how to calculate optimal value for “pool”
parameter in database.yml
I’m using rails 3, ruby 2, postgres, elasticsearch, unicorn, nginx
My server : 12 coeurs 24 threads ( I dont know if it is necessary for
define pool value)
unicorn : worker_processes 24
I have around 1500 pages requested per minutes, and around 200ms per
Unicorn only dispatches one request at a time to each worker, so the
setting doesn’t matter: each worker will only process one request at a
Your numbers suggest you can do 5 requests per second per worker, ie 300
per minute per worker, so it seems that you have ample resources.
you’ll probably need to do some testing to confirm this as it would only
true if CPU is your bottleneck (as opposed to database concurrency for
Is it also useful/necessary to use this ?
not sure what that is without some context.