crunchy-proxy VS go-pg

This is why I always recommend using a postgres proxy so you can point it to different database servers as you stand them up. AWS does this by default but GCP doesn’t yet, so you have to have your own Postgres proxy (e.g https://github.com/CrunchyData/crunchy-proxy) or move to an HA Postgres setup like Patroni that has streaming replication (https://github.com/zalando/patroni).

Might have been this one but I can't be sure https://github.com/go-pg/pg

I usually work with PostgreSQL and have used https://github.com/go-pg/pg. https://github.com/uptrace/bun is its successor.

And it is go-pg. I personally really like it as a query builder (rather than an ORM) and it has good performance. It has one feature which I'll tell you about next, without which I would have struggled to implement what I had in mind. But this feature is available in other libraries too, so don't rush to rewrite your code.

bun and go-pg both seem to work fairly well for these kind of purposes.

For database communication we will use go-pg module. You can install it by running go get github.com/go-pg/pg/v10. This will install version 10 of go-pg module, which is latest version at the time of writing this guide. Now create new directory internal/database/ and file database.go inside of it.

gorm.io/gorm`, or `github.com/go-pg/pg` if you need something more feature-rich. I would avoid using `go-kit` and `go-micro` "frameworks" unless you have some special requirement for it. But you can take a look to have an idea of what Go "framework" may look like.

This one is mostly excellent: https://github.com/go-pg/pg