In the event Log's search for Audit Failure and see if there is a message looking like "Unable to launch from Pool XXXX for user domain\username: No co-management availability for protocol PCoIP"
Next step, look at your connection servers and take a look at your CPU usage.
Lastly, add a few more vCPU to your Connection Server's. This worked for me to alleviate 99% of that issue. Though it still happen's from time to time, i know there is an open Engineering project internal to VMware on the issue.
