The reason is quite trivial: most users want to use HTTP API for monitoring, whether it is the optimal thing to do or not.
We have discussed at least two plugins that can "push" notifications about certain pieces of system somewhere, whether for auditing
or monitoring.We already expose internal events via an exchange in at least two other places: the amq.rabbitm q.log exchange
and the event-exchange plugin.
In the end both ideas are just that, ideas that are never a priority. Specifically for stats, there's an extra problem to overcome: these days, stats are stored
across all cluster nodes (as of 3.6.7) and aggregated by the node that handles HTTP API requests. So it would not be as trivial as the event-exchange
plugin.