Even so, there has still been a lot of interest in GraphQL so users can leverage the growing ecosystem for things reflecting the data model/types for client usage, offline caching, etc
528 karma · joined August 15, 2019
Even so, there has still been a lot of interest in GraphQL so users can leverage the growing ecosystem for things reflecting the data model/types for client usage, offline caching, etc
https://supabase.github.io/pg_graphql/roadmap/
Its early days, so the conversations around aggregates haven't happened yet, but I'm optimistic that they'll make an appearance in a future release
Postgres runs in a 1 GB VM all by itself with some optimizations to get the most that limited hardware
All other services are in a second 1 GB VM.
You can see a list of those services here https://github.com/supabase/supabase/blob/master/docker/dock...
The memory use per process can differs by use-case, so its important to leave a bit of headroom. For example, PostgREST's memory consumption can grow if the amount of data being returned from its queries is large.
By the time you include:
- supabase studio
- kong
- auth
- storage
- meta
if they each take only 100MB, its pretty snug on a 1 GB VM!
pg_graphql: https://supabase.github.io/pg_graphql/performance/
graphile: https://www.graphile.org/postgraphile/performance/
We'll certainly be keeping an eye on performance as it gets closer to GA
as samwillis mentioned, the memory footprint is tiny, which is a big perk for supabase's platform (or if you're self hosting) but its also fully language agnostic which opens up lots of options for extensibility:
For simple use-cases you can expose the graphql functionality over http using a PostgREST as described here https://supabase.github.io/pg_graphql/quickstart/
but, if you want more configuration like adding in middleware or wiring it up to an existing backend application, you can do that from any programming language that can connect to postgres, rather than only javascript
The columns and tables that are visible are also controlled by the role.
One cool thing about that approach is you could run e.g. an admin API and a user facing API all from that same endpoint by executing as different postgres roles!
Yes, it does! Any columns that are not selectable by the user's role are hidden from the JSON `response` and `old_response` keys and are not allowed for user defined filtered.
Here is where the columns are checked to see if they are selectable https://github.com/supabase/walrus/blob/f030688d9d94520b916d...
And here is there they are filtered out https://github.com/supabase/walrus/blob/f030688d9d94520b916d...
Nebulo is very much a work-in-progress, but its reached proof-of-concept viability so I'd love to hear what the community thinks!
At the moment, the motivation for the CLI is mostly to give quick feedback to people who want to try out the project. I'm curious if anyone has a use-case the CLI running a production ready web server?
In the short-term I expect to focus query safety and the last remaining elements of the relay connection spec (reverse pagination) that aren't implemented yet. https://relay.dev/graphql/connections.htm
Please feel free to raise any killer feature ideas in the comments
Contributors wanted!
Shutout to a few related projects written in other languages that inspired development of Nebulo (via envy):
JS: https://www.graphile.org/ Haskell: https://hasura.io/ Clojure: https://github.com/graphqlize/graphqlize lmk if I missed one