single point of failure

[norcis]

I deciding what to use ARI or AGI for high availability scalable application. With AGI I can use web proxy to load balance request. In case of ARI proxy, how to avoid it becoming risk of single point of failure?

[skrusty]

ARI proxy removes that by ensuring multiple application nodes can handle calls.

Your design needs to ensure you don't create bottle necks, for example in your message queue system (have high-availability etc).