streamline log monitoring process

[jjksexton]

Our current log monitoring practices are spotty. The log reports sent to libsys/cdr folder are too numerous and don't do enough to draw our attention to errors when they crop up. How can we improve log reporting so that we are aware of problems when they occur? Can we identify a set of logs and warning levels that are especially important and send direct notifications of those errors to us individually instead of to a shared mail folder? (Similar to the appdev team's nightly php errors report)