I have also been having problems the last 3 days.

I have a linux system with two NVIDIA cards, mining with two instances of rpcminer, each pointed at mining.eligius.st:8337.  I haven't touched the configuration of these miners in several weeks.  Just in the last days, each instance stops finding shares (at different times) until I restart them, yet each instance keeps its GPU busy.  This happens a couple times a day for each instance of rpcminer.  Would it be worthwhile to switch over to mining.eligi.us?  Or has your new DNS become stable?

I also have a new system, with a Radeon 7950.  The system is in flux as I have switched off between Windows 7 and Windows 8.  Having limited time, mining hasn't been my top priority, but I did configure the default Eligius pull-down server option of guiminer (v2012-12-03).  It worked well the first couple of days, giving a fairly consistent 400 MH/s.  But over the last three days, it reports incessant, but intermittent "connection problems".  Eventually, it stops mining until I Stop Mining/Start Mining.  The default Eligius pull-down uses poclbm, pointed at mining.eligius.st:8337.

I'll try BFGminer for the 7950, but that wouldn't explain the similar problems with my Linux system.  The only things common between the two systems are the networks and computers from my ISP to your pool.  It seems likely you have changed something on your end.  Perhaps your new DNS provider or your new DNS configuration?