What is forward rate limiting?
In the context of using "forward rate limiting" at a given edge server, forward rate limiting is used to describe a process by which the edge server is instructed to dynamically adjust the rate at which it sends requests to the origin server.
This rate adjustment may be based on an estimated load being sent and a number of failures or slow connections encountered under that load.
In the absence of forward rate limiting, the edge server attempts as many connections to the origin server as are needed to satisfy client requests. And, it will forward requests as quickly as possible, without regard to the responsiveness of the origin server.