No, that doesn't sound right, wot is calculated concurrently, it shouldn't be holding up feeds (unless too many requests are being sent to a single relay).
> Having a fallback mechanism sounds like the best of both worlds.
Yes, that's what I'm saying