Problem:
I wonder how many hosts should be communicating with 1 foreman-proxy server? Is there a general guidance on that? Is there a host limit which may connect to a foreman-proxy? Can too many foreman-proxy servers make some problems?
Can you share your experience?
Thanks
There is the Performance Tuning guide which contains a good overview and explanation of the tuning profiles. From my experience these match quiet well for the foreman server itself.
On the Smart Proxy I never had any problem requiring tuning. In most environments it was split up into multiple instances not because of performance but network policies or departments. Only the managed service could be a problem when I think about Puppet for example where I perhaps need some more compilers or Pulp which could slow down with multiple parallel access.
Thank you. I wasn’t aware of this page, couldn’t find it. It is really helpful 
One thing to consider: this guide is aimed at “full fledged” systems (= with content, ansible etc features on).
The expected requirements will be much lower if you don’t use certain features.