The space is def interesting and the product probably lowers the barrier to entry for development. My main question will be why would you use it against plain/open-source PostgreSQL RBAC + GraphQL server or something like https://www.graphile.org/postgraphile/ ?
I considered Postgraphile for our company but picked Hasura. While PG has better extensibility, Hasura has a more polished unboxing experience and way lower learning curve. Time to market is crucial for us as we needed to prove our MVP.
That's technically true, though Hasura is a well-funded startup with a $99+/mo cloud offering and Enterprise features while postgraphile is one main open-source developer who has support contracts available and accepts donations (the "pro" package is pretty minimal and costs just ~$25/mo).
You can get started with Hasura for free which is great, and you can run it on your own servers (if you want to manage a Haskell service) which is also great, but in practice choosing Hasura means you're relying on a company for your backend and choosing postgraphile means you're relying on an express plugin which you can get a support plan for.
I'm not saying one is the better choice than the other for this reason, many would prefer the funded company.
Nowadays most of these self-hosted apps run on Docker containers as a wonderful abstraction.
For example, in my company we have self hosted Metabase on App Engine Flex. It is written with Clojure and runs on the JVM. I know nothing about these things, yet I was able to make it run with high availability.
You could also run it on Kubernetes or other similar options elsewhere.
That's a good point; I've always assumed this isn't common with Hasura but I honestly have no idea. $99/mo just seems like a great deal if your bandwidth usage fits within the included 20GB/mo.
IMO both Postgresql RBAC and Graphile make a big difference.
The work that has been done over Graphile has been tremendous, very very polished and well documented.
I have been using it on https://stack.lol with great results.
Edit: I forgot that I was trying a "learning repository" pattern where I put longform comments throughout the code. A little more difficult to discover than markdown, as I've learned.
In general, and for SQL Server specifically, our intent was to add great support (low footprint, works out of the box) in a way that doesn't need DDL or write access to the underlying database.
This becomes super useful for folks building new applications or new features against existing SQL Server systems (which is a rather large ecosystem beyond just the database, since so many products use SQL Server underneath too!)
