Major Description
As a safety net, it would be helpful to rate limit the number of rest requests/sec that we accept/process at a time by the cluster manager/ns_server to avoid overloading ns_server process in-terms of both memory and CPU.
It would make sense to have a high-level rate limit as a backstop against bad actors or misconfigured clients / services. This is meant as a backstop / safety net for ns_server. It is not considered a full solution to rate limiting for the product itself.
Remaining Questions:
What is a default rate that makes sense? Should it be enabled by default? Is this necessary at all (or should it be replaced by something more comprehensive)?