Github user phunt commented on the issue: https://github.com/apache/zookeeper/pull/440 I like the idea of moving to dropwizard vs our own homebrew. Seems like it would allow us richer insight for metrics. Here are my concerns: 1) performance. Do we know the impact of this library vs our homebrew? Both in the small and the large. 2) likely this change will go into master (3.6.0+) - impact on backports. I'm assuming we would want to take advantage of this library across the board. (can we do that? I assume yes?) 3) I believe I recommended on the original PR separating the new metric from introducing new dependencies. I still believe that's the quickest way to get the feature addressed, vs introducing new metrics functionality. Can we commit #415 independently of this change?
---