Just to clarify, what I wrote at point 3. was about Attended concurrent.
Now let me explain a bit about Unattended:
As you can see in the attached image I have 3 machines, 2 Windows Servers with 15 robots each (ROQAWRDS20,27) and one Windows 10 machine, 1 robot (ANOPREA). Once a robot becomes available in Orchestrator it will consume 1 runtime and License status will be displayed with green. Once the maximum number of runtimes is exceeded all other new robots will be available, but the License Status will be displayed with orange, meaning that the machine is not active. In order to make an inactive machine, active, you’ll have to free a runtime (you can only do this by disconnecting an active robot or killing robot service). As you can see in my ss, starting with 18.4 we’ve added a toggle with which you can deactivate a machine in order to free runtimes for other machine(s).
The concurrency for Unattended robots is possible in case of Windows Servers with multiple robots (different usersname, same machine). In my example I have 15 robots on two Windows Servers for which I allocated 15 respectively 10 runtimes. For ROQAWRDS27 this means that only 10 robots can get licensed at a time. Saying that I start a Job on all 15 robots from ROQAWRDS27, 10 robots will get licensed and will run the job, 5 will remain in pending until the first 10 finish and then they can get licensed and run the job.
In case of ROQAWRDS20, you can see that even though I allocated 15 runtimes to that machine, at the time of my ss only 1 robot was licensed (you can see that in the “Used” column).
So, for now, this is intended behavior, not a bug.
Please let me know if you have more questions,