Hi Paneendra,
I believe some of those items, if not all, may be discussed in the CA Release Automation 5.x / 6.x Architecture and Implementation Best Practice Guide.
Items 3 and 4 look the same to me. But I would imagine that most of the restrictions related to simultaneous deployments are going to be resource based. Not just memory and/or cpu which is often looked at quite closely. Disk I/O bottlenecks are sometimes overlooked and not just on the management server. They should be considered NES, NAG, as well as DB (very important). For trying to find the appropriate balance/scale of what you're looking for you might consider involving system administration teams to get their recommendations on monitoring these facets of server/application performance.
Kind regards,
Gregg