News for hackers. Stuff that matters.
Re:Streaming data in Postgres to 1M clients with GraphQL by atentaten on Monday October 17, 2022 @08:58AM (#33232826)
How different is this from Supabase Realtime?
Re:Streaming data in Postgres to 1M clients with GraphQL by tango12 on Monday October 17, 2022 @09:12AM (#33232993)
Haven’t looked at the internals or the API, but I think for starters, Hasura is designed for working on any Postgres including ones that have existing data, especially modern Postgres offerings (crunchy, alloy db, neon, Aurora, cosmos Postgres, timescale, yugabyte)
Re:Streaming data in Postgres to 1M clients with GraphQL by Dystopian on Monday October 17, 2022 @10:24AM (#33233909)
Disclaimer - I work on Hasura, so I won't comment too much on which is better.

They're all similar flavors of producing realtime results - which take similar, but different, methods to their approach.

My understanding (please feel free to correct me if I'm wrong):

- Supabase Realtime uses WAL.

- Hasura Streaming Subscriptions uses a query which will be append-only (could be a sort-by or also WAL).

- Hasura Live Queries uses interval polling, refetching, and multiplexing.

- Supabase uses Postgres RLS for authorization, while Hasura uses an internal RLS system which composes queries (which allows for features like the multiplexing above).

- All 3 use websockets for their client communication.

Supabase Realtime

https://github.com/supabase/realtime#introduction

https://supabase.com/docs/guides/realtime

Hasura Streaming Subscriptions / Live Queries

https://github.com/hasura/graphql-engine/blob/master/archite...

https://github.com/hasura/graphql-engine/blob/master/archite...

Re:Streaming data in Postgres to 1M clients with GraphQL by atentaten on Monday October 17, 2022 @11:59AM (#33235423)
Thanks for this. I was more interested with the differences in implementation rather than which one is better.