You're probably using the wrong thread group for what you are trying to do.
The concurencly thread group ramps up the load until the number of concurrectn requests being made is (in your case) 2.
Given your requests are < 0.5 msecs seach (eg response times are all 0ms, and a throughput of 2000/sec)
That will mean you NEED ~4000 incoming calls per second to supply load that results in an average concurrency of ~2. (2 / 0.0005)
If your threads have delays, then this might require 10s of thousands of threads to achieve!
It comes down to what your testing goals; I suggest moving to a rate-based model, eg "visits per second" or "logins per second" which is more reasonable and understandable at a businss level.
I suspect your sytem doesn't need to handle 4000+ calls per second :-)