Another day, another fail-safe.
Not sure I appreciate your tone, sir.

Last fail-safe was 2014-05-15, 12 days ago. My code updates have made this a much less common occurrence than in the past.
The problem is that when it does happen, if I'm not able to correct it quickly the backlog of shares it needs to process to catch up grows and it takes much longer to recover.

I'm testing some new code that will allow the system to continue with auto payouts and balance updates and all during a failsafe, but doing so with limited capacity (holding some payouts to those potentially effected by a payout in an orphaned block's coinbase, for example) until I'm able to do my normal manual look-over of everything. This should also make catching up much quicker, since it will never really have fallen behind.
This code is right behind the new and secure NMC payout code, which has been giving me a few headaches lately, but is very near completion.
What events can cause the CPPSRB reward system to enter fail-safe mode ?
Many things, actually... the most common cause: a competing chain race that lasts more than a predetermined amount of time which results in Eligius being on the losing side of the chain (orphaned block that CPPSRB has no way of knowing about until after it has processed quite a few shares afterward). Others include some internal sanity check not passing when calculating the coinbase payout list and other critical items.