As I get news from past, A default ALM server (Pre-requisite Configuration mentioned), can handle 200 concurrent users at a time, any load beyond that, should route to the Node B or Node 2 (If Cluster Env).
I think this is the threshold limit what HP suggests.
I am not quite supportive over this, I have checked the application more than 250 users at a time in my ENV and nothing happend.
Any ways, it allways depend upon the Heap Size/Memory.