Around 1600 CST on 2020-11-26 we observed that approximately 20% of the time the api server at https://acme-v02.api.letsencrypt.org/ will stall after ack'ing the client-hello when queried using curl. We observed this from multiple locations and network providers on IPv4 and IPv6. Next packet should be a server-hello but this is never sent.
This site is fronted by cloudflare and the intermittent but repeatable nature would point to a possible dead-node behind a load balancer which is periodically being round-robin'd and failing.
Thanks for the report! Can you confirm whether this was 1600 CST or 1600 UTC, and whether you're still observing timeouts? We've received a few other reports and have been investigating.