and i think we could/should almost definitely lift the

UrlToFetch -> Snapshot

part into python code then, which uses a lower-level intrinsic for http requests internally