In the AMEX lab environment we saw a strange problem, They had a 'tcserver instance' resource marked as unavailable (we checked the tcserver instance and it was running fine) but Hyperic server was still showing metrics for this resource (if it was really unavailable the server cannot get those metrics) and all the tcserver instance services were marked as available.
Restarting the Hyperic server solved this issue. Scott mentioned that he has seen this issue before and that this is probably a problem in the Availability Manager. In the Sofia scale lab this happens a lot of times with a lot of different resources types, might be related to the Availability cache, should consider cleaning this cache every ~1 hour because this issue happens after the server is up for a few hours and it seems that cleaning the Availability cache solves this. Cleaning the cache every hour probably won't affect the performance too much.