16GB_of_ram

Becoming more like an RDS alternative and Planet Scale competitor is good to see.

Where do you host it?

How is this different from grafana?

it was down on paid and enterprise plans too

Cloudflare SQL or Durable Objects + Custom auth + Cloudflare workers + R2. Better Auth. NO monthly fee, LARGER FREE TIER. Little more setup, but much easier than self hosting.

No, it doesn’t. You can have a Cloudflare worker serve as middleware in between defining what what folders are public versus private in a single private bucket

How did you manage the high api influx? what compute are you running? Thx

That's why you add logs in edge functions. And by the way how are you even doing push notifications without an FCM token or Expo Notification Id (I think thats what expo has). I recommend using FCM tokens with edge functions. I used edge functions to send more than 600k notifications this is NOT a supabase issue.

only the web ui goes down sometimes other than that the api only goes down when things like cloudflare or aws go down

got you

It fully depends on what you're doing. For my mobile apps supabase. However, this past month I've been working on dirty cheap session replay that is self-hostable and open source for react native, so that needs a lot more customization.

Express JS on Node, Fast API, Zod, Better Auth, and Drizzle PG is all you need on a VPS 100%

If you are running an app, supabase is not just a postgressql but everything you ever need. No other thrid party auth, edge functions, etc. If you are running like a website postgressql on a VPS with something like better auth makes more sense. Think of it like this: Supabase is AWS. It has a bunch of micro services that are well suited for an app back end.

I've been on supabase for a few years all the outages are stuff like dashboard not working once every year for like 30 minutes. No biggie. The only big stuff where users are affected occur when AWS and cloudflare fail.

I've been on supabase for a few years all the outages are stuff like dashboard not working once every year for like 30 minutes. No biggie. The only big stuff where users are affected occur when AWS and cloudflare fail.

trust me all the outages are due to upstream providers which if you choose another cloud you would have been affected by too anyway.

trust me all the outages are due to upstream providers which if you choose another cloud you would have been affected by too anyway.