Successes and limitations of New Keynesian DSGE models?

What are some good review papers that discuss the successes and limitations of current state of the art (New Keynesian) DSGE models assuming rational expectations? Probably some of these reviews can be found in the boundedly rational expectations (RE) literature. Especially a comparison to some (S)VAR results and some well established patterns in the macrolevel data would be nice. The limitations of RE as such are well known, but to me it’s unknown what the difficulties are with models assuming RE in explaining the data.