Here is the documentation page on this page:
https://docs.mashery.com/design/GUID-586E5F93-7538-42C1-990E-E88DADAABA24.html
 

This article provides answers to some frequently asked questions on this topic.

Q1) What is simple round robin mechanism?
A1) As indicated it is a simple-Round-Robin-load-balancing.
Each listed domain is tried for each request, one after the other.
If a backend is down, the proxy_worker would try the next one in the list.

Q2) If I configure 3 backend ip addresses and one of them is down, is a request not sent to that backend?
A2) Yes. The proxy_worker will know its down only when it tries to send the request.

Q3) How frequently does health check happen for the backend ip addresses to identify that its healthy?
A3) As it is a 'simple-round-robin' there is no health check that happens. The proxy_worker will not know it is down until it tries to send a request to that backend.

Q4) If the proxy_worker tried the second backend and it was down , will it move to the next one? If next one is also down, will it come back to the first one or will it give error to upstream with 504 error if the last one in the list is also not available?
A4) Yes. If the proxy_worker is not able to connect any domain in the list, a 504 is returned to the client. If 504's continue then failsafe kicks in.
More details about failsafe : https://support.tibco.com/s/article/Fail-Safe