I think you will face a challenge here, as this relates to some backend server measurement against front end user behaviours:
Concurrent users on your web server relates to typical Concurrent user nodes the server allows at max at any one time. So lets say a max of 25000 users concurrently are on your server, then user 25001 tries to access site he would reach server busy message. This is a limit at any one time at once.
In analytics the closest I recommend you use to that metric is unique visitor. You can then break it down by hour and make a close assumption to your IT team its close to concurrent user max. It will be lower than their limit as this is an hourly breakout.