We have seen this lately when someone ''changes'' the BASE_REPOSITORY_PATH in the Site Configuration parameters from the Site Configuration tab while trying to point QC to a DIFFERENT repository path.
There was another EyeOnTesting post recently where someone changed the path and it was not fully defined (now in 12.x, QC expects the whole repository-related structure to be there).

http://eyeontesting.com/questions/5600/getting-server-muted-error-after-changing-base-rep.html

Another way this can happen is if you are using a network shared location for your repository (perhaps as part of an ALM server cluster) and something changes in the way your QC Service user gets to that location (password expired, alias domain changes, other network changes).

In your case, it sounds like no EXPLICIT change was done to the path, but this ''just happened'' after a service cycle leading me to believe that this was a ''system'' change behind the scenes.

Check the value of BASE_REPOSITORY_PATH row in the PARAMS table in the qcsiteadmin_db schema/database used by your instance of QC. Confirm the path exists and that it has the SAME structure and files as the ORIGINAL one (copying all but the actual project repository folders from ORIGINAL is a good start). Also, if a shared location (\machineNamepath) check that it is still valid for the service user to use -- some companies will ''retire'' and alias from time to time.

I have also seen this same error situation in the instance where the free disk space and/or available memory has been depleted as well. Just something else to be made aware of and pay attention as for a possible source of this error condition.