the pool is great and tycho you do a good job always monitoring and getting it back up as quickly as you can and you respond right away. but that doesn't discount the fact that you keep going down reguarly, pretty much almost daily now, and for extended periods of time. can you setup something so that there is 0 downtime? an automatic backup of some sorts?
just run two instances of miner per gpu, each with different priority, higher for your favourite pool and lower for backup one
if main pool is down all work is done in second instance and submitted to backup pool:
ie:
poclbm.exe --verbose --vectors --worksize=128 --frames=60 --host=mining.bitcoin.cz --port=8332 --user=m4rkiz.0 --pass=XX --device=0 -r 3
poclbm.exe --verbose --vectors --worksize=128 --frames=1 --host=deepbit.net --port=8332 --user=m4rkiz@XXXX.XX_0 --pass=XX --device=0 -r 3
with this setup deepbit is your main pool and slush has around 1 MHps until deepbit is down