So far I've counted 6 timeouts (including one cancelled) which are
wasteful of computing resources. I will re-enable network timeout
in the manifest for the moment.
More importantly, we need to figure out why the latency got so horrible
with 2 pinning services (Pinata and Estuary) and AFAIK 2 local nodes,
especially after go-ipfs introduces NAT hole punching (I enable it for
my node at least).