_ihm40

I might reccomend either opening a Github discussion or submitting a feature request. Reddit can be a difficult place to get visibility on things like this

Yeah in practise this no different from a user that signs up via a magic link. This is also why a user that is already logged in and hasn't gone through password reset can update their password but in those situations you might confirm their old password first

I think practically if you need >400s to perform some long running operation then yes you have to offload that work to some other service like AWS lambda functions or Netlify (i've never used this so can't attest to it's suitability)

What is the naming convention for the full list of migrations in supabase/migrations? I'm guessing alongside timestamp it would reference module\_x for the module the migration belongs to? Are there any module specific tables that might conflict with other things? Have you considered having a schema for each module?

Anecdotally i have seen that more people use separate projects rather than a persistent branch for staging

The first thing is that you can probably read customer stories of teams using Supabase here https://supabase.com/customers. This might be the most helpful resource about understanding the experience of larger teams on Supabase. When it comes to individual components then since the Database is Postgres that can obviously take you quite far. I have never seen anyone complain about auth/storage issues purely as a function of scaling. People have highlighted a few times that realtime can become more expensive as it scales but i don't use realtime enough to verify that claim and sometimes people want more flexibility from server side operations that edge functions can provide due to it's limits https://supabase.com/docs/guides/functions/limits but i think you can probably make a lot of optimisations to the code before that is an issue.

Perhaps it is worth making a feature request on Supabase https://github.com/orgs/supabase/discussions/categories/feature-requests?

I don't believe there is anything exactly like that feature unfortunately but destructive queries does tend to give the following warning https://preview.redd.it/ttf6f89p0gkg1.png?width=1162&format=png&auto=webp&s=7a011057a9399a5f90388666ec24c2e1374a7881 which is helpful for double checking purposes

Hey there! Norwegian will likely make it difficult to find people who can respond to your query but judging from google translate you would like some paid help to assist you using Supabase & Lovable. I would suggest going to our discord server here [https://discord.gg/YJWZBD23](https://discord.gg/YJWZBD23) and adding a request (in english) to the #jobs channel

Hey there, yeah there were some DNS issues related to users in the United States. It seems to have been resolved for now but you can check [https://status.supabase.com/](https://status.supabase.com/)

Yeah that is strange, haven't seen that before.

Which guide were you following, i saw that clerk updated this 3 days ago [https://clerk.com/docs/guides/development/integrations/databases/supabase](https://clerk.com/docs/guides/development/integrations/databases/supabase) so may have more up to date info?

This may be the doc being referenced [https://supabase.com/docs/guides/storage/s3/authentication](https://supabase.com/docs/guides/storage/s3/authentication) but it doesn't seeem like you can use the supabase client to send to S3 but instead use the S3 client to place objects in your supabase storage. I think if you want to place files in S3 then i think you would have to use the aws client.

I wouldn't claim this is "best practise" but most people who want to isolate their databases i've noticed have usually done dev changes locally with migration files, pushed those dev changes to a supabse project which is the staging environment and then if everything looks good use the cli to push migration changes to prod.

I don't think these are necessarily anything to worry about as there are probably some background processes supabase as a platform is doing with your database that causes it to produce logs. For example, i would assume that the dashboard/table UI requires connecting to the underlying postgres to display the information and those connections would produce logs which may be what you are seeing here.

If the user signs out or clears browsing data then you won't be able to re-create that same user because this idea of 'logs out' and 'logs back in' doesn't seem to be the intended purpose of sign in anonymously. It is only used for accounts which are temporary or throw away by nature such as shopping carts. You don't necessarily have to log them out of your game but if they clear browser storage then they won't be able to access information related to them.

if you look here [https://supabase.com/docs/guides/auth/debugging/error-codes](https://supabase.com/docs/guides/auth/debugging/error-codes) it says that it should provide \`email\_exists\` error code. If you log your data or your error i'm sure it would be present in one of those

Difficult to know just from a 500 error. If you check your auth logs do you have any additional details or messages that can narrow down the issue?

The is less a workaround but more how edge functions should be handled going forward if you have migrated to JWT signing keys. This guide here should explain how securing edge functions can work in this new paradigm https://supabase.com/docs/guides/functions/auth. It should have given you a warning to update your edge functions before you completely migrated over?

When you say Admin user here, do you mean an admin user inside your users table or Admin of your supabase org? I can't recall anything directly for your case but the mcp docs do provide some guidance on using LLM (mainly for development) https://supabase.com/docs/guides/getting-started/mcp. I'm sure it's advice might help in other places also

I have come across this article [https://medium.com/@esme\_35590/how-i-set-up-supabases-inviteuserbyemail-6f7072a0c587](https://medium.com/@esme_35590/how-i-set-up-supabases-inviteuserbyemail-6f7072a0c587) (no affiliation) but it it shows the general gist of what i was thinking which is you invite the user by email, redirect them to a password setup page (this should create a session and log users in) and from there you use supabase.auth.updateUser() to update the password that they provide. It works similar to otp but hopefully without the link expiry.

I think this may be different from plain otp link expiration, if you just invite them by email without specifying it is an otp then it shouldn't expire but create a user for them anyway?

Have you looked into this [https://supabase.com/docs/reference/javascript/auth-admin-inviteuserbyemail](https://supabase.com/docs/reference/javascript/auth-admin-inviteuserbyemail) where ask for them email only, send the invite link and then ask them to fill in the password afterwards

I don't think the preview bot runs if nothing in \`supabase\` folder changes. Does closing the supabase branch and re opening it pointing at the same github branch fix the branch issues