node-pg-migrate VS postgres-meta

Why PG? For the clarity of the article to a large circle of developers, the entire explanation will be built by PostgreSQL and PG package. And in a real project, the database schema will change over time, and in order to be able to perform migrations, we will use Node PG migrate.

b) How you will manage schema migrations: every time you change tables, columns, foreign keys, every time you create or modify stored procedures, and so on, you need to do this through a migration. Here I'm at loss to recommend anything specific, but if you went with Prisma, I would read their docs and use whatever they are recommending you (if they actually recommend something). If you want something more neutral but still in the node ecosystem, node-pg-migrate might be the winner here? I'm not sure.

You’ll probably want something like this too: https://github.com/salsita/node-pg-migrate

You can use any normal DB migration tool. For k8s, I put the app's readiness probe to false, run the migrations and then toggle the probe back to true.

Here are some migration libraries:

Go - https://github.com/golang-migrate/migrate

Node - https://github.com/salsita/node-pg-migrate

What grinds my gears it that someone actually started this process with a simple pull request: ADD CRUD permissions. Someone from Supabase said, "sure, no problem, we will work on this next week"... and nope, nothing. Then the PR was randomly closed yesterday! Why? No idea, no explanation. Granted, this was a simple PR and they could be rethinking an easier way to do this. But I digress.

I picked the first issue, which was a bug in the Supabase dashboard, where users could not edit a database column from a primary type to its array type. The dashboard was built with Nextjs and TailwindCSS, using the Supabase-UI components libraries. It looked like was a quick fix to me. After investigation, the error was not with the dashboard but it was an error of insufficient type handling of column type identifiers within the Posgres-Meta package, which was also written by Supabase Team. Then, I headed to the package upstream and fixed it. The PR got merged and released soon after.