I don't know about the original poster but I have seen this issue quite a bit lately as well with Horizon View Desktops not RDSH. When using the Blast protocol occasionally the user will get the error "The display protocol for this desktop is currently not available error.' The solution is to restart the blast service on the desktop (sometimes works) or restart the desktop completely. This is definitely an issue with the Horizon View agent, but we do seem to notice this happens more often when we have some sort of external connection disruption like flipping the external firewall to a different node. It also seems to happen more often on persistent desktops in our environment. We think because our persistent desktops are not logged off except for patching once a month whereas our non-persistent desktops get force logged off every other week.
My hypothesis is that after X number of connection disruptions the Blast service in the agent doesn't clear the old connections properly and a memory limit of some sort is exhausted and then any new connection can't allocate the memory needed for the display protocol to work properly and it is just dead. This is purely just speculation though.
