We've had two problems with the pool since Friday - we are starting to push the design limits of some of these parts of the system.  On Friday we had a brief outage that was quickly fixed by Eleuthria, related to hitting the upper limits for shares accepted for a single worker.  Another issue cropped up Saturday, in which a portion of the rigs were getting an error returned by the stratum server, which wasn't coming from the pool.  This issue seems to have been related to share difficulty returned through stratum(s) to the rigs.  We've alleviated the problem by balancing the rigs/hashrate across workers, while also managing worker diff settings.
...