Hi @koksime-yap_5909 ,
I'm afraid we can't provide much clearer guidance than that, as there are so many factors involved that make predicting performance basically impossible. For example, the feed types you're using, your CI server configuration, how often developers are rebuilding, etc.
The article you found is actually what we send users who experience symptoms of server overload, to help understand where it comes from and how to prevent it. As the article mentions, the biggest bottleneck is network traffic during peak traffic - there's only so much that a single network card can handle, and scaling CPU/RAM doesn't really help.
This is where load-balancing comes in. The main downside is complexity/cost, which is why a most customers start with a single instance. It can take quite a while for a tool like ProGet to be fully onboarded across teams, so performance problems likely won't happen at first.
Hope that helps, let us know if you have any other questions!
Thanks,
Steve