SLO/SLA for Uplink API and/or other fallbacks


I’m investigating using Apollo Studio with a few members of my team for our company. We are planning on using Apollo Studio’s managed federation. We have a running gateway which is using the uplink option to load the supergraph on startup instead of polling every graphql server that we run.

One worry we have is what will happen if the uplink API is down. I scoured the docs but couldn’t find any guarantees on the SLA/SLO of that endpoint. In the event it is down, is there a way for the gateway to fallback to a file or some other (potentially outdated) graph?

Due to the nature of our daily traffic, we automatically scale instances to meet demand, so it’s possible for us to create lots of new gateway instances within a short period of time. If the uplink is down, it is my understanding that the instances would be totally un-usable until it came back.

Thanks so much,