Neon has pg18 ready to use on the day it's released, just like last year.
Sadly one of the most interesting features, Async I/O, is going to take some work at the file system level for us to support. For now all operations are still sync. And we'll keep updating on our progress there.
cold starts are 500ms on average, and that's only for the first call that wakes up the db from hibernation. people still seem to think that this latency happens for every call (see other threads here) but once the service has woken up (cold start over) you're back to regular (sub 10ms) latency timings and the service continues to run that way. you'll only hit a cold start again if (you have this option turned on) your service goes idle for > 5 min. You can turn scale-to-zero off and you'll run 24/7, have zero cold starts.
$19 plan is going away, will launch a better $5 plan soon.
there isn't something like Vitess for Postgres yet but there needs to be. Migrations are painful in general and they become very painful at scale. i haven't used gel yet, i know it manages migrations but i don't know to what extent. most of my experience is with prisma, drizzle, and atlas.
neon is working on some plans to solve migrations at scale. we think our custom storage layer will allow us to optimize certain paths, like setting a default value in a new column for a table with millions of rows. this alter command can take a lock on a table for hours. but ultimately we need better tooling.
ideally there is a client and a hosted service such that you can use the client to run migrations on your own from the CLI and integrate it into your dev workflow. the hosted service version allows you to push up your schema change from the client to an API. from there you can manage the migration rollout from an operational dashboard that helps you tune resourcing.
when i was at github we used vitess to roll out a migration that took 3 weeks to complete. a long time to wait but that's a better tradeoff compared to a migration that takes down production for 6 hours.
we have multi-region disaster recovery and replicas coming in 2025. Switch over time will be greater than this active-active system but the overall latency for Neon should be much less on a consistent basis.
Sadly one of the most interesting features, Async I/O, is going to take some work at the file system level for us to support. For now all operations are still sync. And we'll keep updating on our progress there.