Yes, there are use cases for high performance JS. We have JS that we use client side that we also use server side for Hadoop processing of data. So we didn't w

Yes, there are use cases for high performance JS. We have JS that we use client side that we also use server side for Hadoop processing of data. So we didn't want to rewrite and maintain this large codebase in both Scala/Python and JS, so we ran Node on Hadoop. No, it really isn't setting yourself up for failure if you do it correctly. Node really is quite fast; for our use case it rivaled Java 8 runtime in MR CPU time, the only downside is it had double the memory footprint. With enough hardware, memory is not too much of an issue.