Add config switch that will dump performance data to rails log at warn level if threshold is exceeded

The HyperModel fetch serverop dumps a nice performance log to the rails log, however its at debug level.

Add a config switch `warn_if_fetch_takes_longer_than` default 0.1 (?)

If the time is exceeded then the info is dumped at the warn level so it will show up in a production log, which can help in diagnosing some data base problems.