Anyone have any real world recommendations for what's required for foreman
deployments of various scales?
So 100-1000 nodes, how much RAM, CPU, etc.
1000-10000 nodes?
Probably looking in around the 2000-5000 node range and wondering what kind
of hardware we'd need if we were splitting that amongst say 5
smart-proxies, and a single foreman instance also running as puppet master
and puppet CA.
That depends heavily on how often you run puppet, if you have scripts
that make API calls, how many users will use it…
Just a single foreman instance bit worrisome, at least 2 behind a load
balancer would provide you a backup in case of HW failure, security
could be compromised in your host, etc… It's best to have this
automated (the puppet-foreman module can help) so that you can
experiment with different box sizes. We do have minimum requirements you
can check here:
https://theforeman.org/manuals/1.12/index.html#3.1.2HardwareRequirements
We don't have a 'reference architecture', however as for real world
recommendations go, you can take a look at blog posts from people from
our community like Chris Pisano who recently wrote about making a high
availability setup in Foreman:
https://theforeman.org/2015/12/journey_to_high_availability.html
Our youtube channel also contains case studies that could be interesting
for that - https://www.youtube.com/channel/UCCo7AZ1oG6TbG0-dwjRqCmw -
and 'real world' users hang out in #theforeman IRC channel all the time,
so feel free to jump in and ask any time there too.
Let me know if this was helpful, sorry if I couldn't be very precise but
it's hard to with little information about the actual deployment 
–
Daniel Lobato Garcia
@dLobatog
blog.daniellobato.me
daniellobato.me
GPG: http://keys.gnupg.net/pks/lookup?op=get&search=0x7A92D6DD38D6DE30
Keybase: https://keybase.io/elobato