I understand this is a loaded question, however I am going to try my luck to see if anyone has info/documentation that I have been unable to locate thus far. Perhaps someone with a better understanding of REST API functionality could point me in the correct direction.
In looking to deploy Remedy 9.1, I am being told REST API will be disabled due to performance concerns of the front end application itself (web interface). I am trying to find out if there is any quality control or prioritization that occurs on the backend that would mitigate this concern.
I understand that there are some obvious savings in not having to dynamically render a webpage or engage the front end virtually at all when making a REST API call, so when pulling data its more lean 1 to 1 to use REST. However if someone where to be reckless with a REST API call, is the ARServer at all equipped to manage this request by assigning it low priority, or would it simply take down the whole system?
I think the bigger issue that is being avoided is that there are kinds of places for bottlenecks. If they are concerned about that, what guardrails do they have in place to protect from an overloaded database? Or a table lock? The REST API is just another client. There is only so much you can do to protect the system. Since it is another client, it has all the protections at the app level that exist for every other client. Indexing, requiring search criteria, threading, and other performance tactics will help regardless of the client.
RCJ has a good point in that you can dedicate servers in the server group (or even outside of it) to accommodate their needs. But you will always go back to the central database as the ultimate risk.