I built https://github.com/k0nserv/plid with Pgrx and had a great time. I did have to scale back some of the magic (dropping derive PostgresType etc), but even so the support pgrx provides is excellent. I also talked to the maintainers a bit in discord and they were super helpful.
The one downside of custom extensions is that you aren’t, AFAIK, able to use them with many hosted Postgres installs, notably AWS RDS.
Maybe one of the reasons why hosted postgres often disallows extensions is due to security concerns from loading arbitrary machine code on a shared host. I wonder if pgrx changes the calculus here.
Since it's a procedural language, you can't do things like create a new index implementation or something else super low level. But there's still a lot you _can_ do. Like implement a custom comparator for a custom type and then use that type in a btree index.
Supabase is an interesting middle ground here — it runs on managed Postgres but gives you access to a curated set of extensions (pg_cron, pgvector, unaccent, PostGIS etc.) without needing to build your own. We used unaccent + GIN indexes for fuzzy city search and it worked well. Still not the same as arbitrary custom extensions, but covers a lot of practical use cases that RDS won't touch.
If I need to do fuzzy location searches, I've thought about just matching geohash locations... even if doing multiple sets and collating them... Note: this was in consideration for something like Scylla or Dynamo.
[0] https://github.com/postgresml/postgresml
The one downside of custom extensions is that you aren’t, AFAIK, able to use them with many hosted Postgres installs, notably AWS RDS.
Since it's a procedural language, you can't do things like create a new index implementation or something else super low level. But there's still a lot you _can_ do. Like implement a custom comparator for a custom type and then use that type in a btree index.
https://docs.aws.amazon.com/AmazonRDS/latest/PostgreSQLRelea...
Which includes literally all 4 extensions you’ve mentioned.
Every managed service does this, specifically because they need to blacklist extensions that touch on the managed parts of it — eg filesystem
Number of memory unsafety and race conditions I've had to debug in production in a year of use: zero.
pgrx is fantastic.