Logging improvements

[jwarkentin]

I've still been thinking about the log filtering problem discussed here. After spending a lot of time recently using Log4J with Kotlin I think there are some important design concepts there. In particular I'm thinking about the way appenders and filters work, the context-based cascading and the declarative config for managing it all.

Importantly, this can't be complex like Log4J. It needs to work out of the box, be simple to understand and configure and allow enough flexibility to be generally useful. It should also be able to improve logging performance when running in production mode, not hurt it.

It might be worth re-thinking Node Monkey as a true logging system first and foremost, with one appender being a browser console appender or something. I'm not sure how the SSH stuff would fit into this but it probably doesn't. Most likely, that will have to move to its own project.