frozenicechan

eu-west-2 is working for me as well just fine, no issues reported on cloudflare's front either. could be a small set of servers

You'll need to open up a support request to do so as this is mostly a user helping user forum. Is this on a free plan or pro?

One of the main selling points of docker is what's inside always work no matter where it runs, it really doesn't care what's on the outside. Thus select any version you want and 99.9% of the time it'll work as long as you're not do something unique with OS fun which you and most people will not. Personally I would go with the version at 17.6.1.063 (just upgraded myself and its working great) that way you're matching what's the latest in the cloud at the moment. Which you should do the upgrade for your own cloud in Supabase to this version as well while you're at it.

<@1436739289604952094> What's the status of your project say? Have you restarted the project?

It does actually install the CLI within the project. Once added you can run all the commands straight from the project folder. The question is, what specifically are you trying to do? There's can be different tools to setup along with depending on what you're looking for.

Thinking you mean Bolt?

true

Well yeah that would explain why nothing works

When you click unhealth status what does it say is unhealthy?

It right there says status unhealthy. So I would restart the project. Very bottom button on side bar aka **Settings**-> **General **-> **Project Availability** -> **Restart project**

That domain is 100% not available from like 6 different DNS servers.

Sounds fine to me overall. Though if not set on the 16 character ID you could skip the hashing if you don't mind bigger ID. I'd just use the UUID personally. Though that isn't as nice to view but all up to you.

What's the version of the firefox developer edition?

I would try a different browser then if the ones its generating also say not secure enough. That doesn't make a lot of sense lol

Look at the doctor's table do you have any columns using digest?

Wait I got it I think because they mentioned "pgcrypto" directly and doctor data. So they are working with highly sensitive data. 1. Are you sure pgcrypto is available everywhere or just a specific schema? 2. Try adding a explicit cast in your calling like: digest('data', 'sha256'::text)

Edit: I'm confused on the actual doctors/digest piece but I'll ignore that for now. The error message is suggesting it got two pieces of data, some text and something unknown. Are you sure the data being sent is in the correct structure you're intending it to be?

Is there any other service you use with the github login? Maybe go there and make sure that still works for you

Correct, so I would try any other possible way to get around your connect/browsers until hopefully one works

Seen other posts in here about it in the past but no official solutions I've noticed, just it fixes itself overtime. Could try a VPN connection if possible to maybe get around it? Hrmmm and you cleared out both cookies and local storage? If you go to github right now and log out/log back in then go to supabase does that change anything?

What part is currently failing? As in you get logged in but don't see anything? The page just refreshes? Something else?

If you go to Settings -> Infrastructure is there a service version update waiting for you? I actually have one to bring me up to 17.6.1.063 but just haven't done that yet. but all these docker images are available to get including 17.4: https://gallery.ecr.aws/supabase/postgres

My docker and cloud are both on 17.6

We actually already do this similarly for testing purposes so that a dev/qa can jump right in past auth and have a fully functional account without having to go the normal oauth path intended to be used. However this is all locally run so in theory you'd need to have like a function that calls the service account to create the user upon their behalf. aka: auth.admin.createUser https://supabase.com/docs/reference/javascript/auth-admin-createuser As for is this a good idea is debate-able for so many security reasons. But ignoring that I'm sure you can make something work where it makes a random ID and password handed out to the user. Server side it'd still be an email address. Then on the login screen you just append the @email address to the login attempt. User doesn't see the email address at all but effectively you're doing an email login still.

It's still your bank that controls most of it so call them first no matter what. I had an issue where none of my payments worked but it was because their transactions were going through their HQ in a small EU country and so both my banks were blocking the transaction as it was seen as unusual. Remembering now, years ago I helped handle payments for a software company for a little bit and all the time we had EU credit cards declined for US purchases just because their banks were being overly protective.

I would find it unlikely Stripe of all payment processors are the ones blocking the transaction but you never know. Thus why I would reach out to your bank if you haven't.

If you haven't yet I would call up your bank and ask if they are receiving the attempts of you trying to purchase. If they have not then it's something with Supabase's payment processor most likely, however if its your bank blocking the transaction for some reason then you can sort it out through them.

Also if you do a custom report of Disk IO remaining under the Observability dashboard what does it show?

Can you share the email/dashboard notify (without any personal details)? Those graphs don't suggest any heavy usage at all like you're thinking.

Per a mod earlier, it doesn't look possible: https://discord.com/channels/839993398554656828/1471459372906446974/1471471962181337265

At this point all you can do is contact support directly and ask if there's a chance you can get it back. Learn what happened with the usage and prevent it from happening. Personally I feel however since this is a Free Project I doubt they'll do much unless you end up paying for Pro plan immediately. Especially if the project disappeared this quickly.

Well it also says as long as you lower your usage within your grace period. However it looks like you didn't or went significantly over so the system stopped the problem.

A bit confused on your messaging due to the error saying it doesn't exist but also saying they are visible. So you see it under the cron table in the database? Have you tried deleting the job from the cron table by its id or is that what caused the problem to begin with? What if you do something like: ```SELECT cron.unschedule('your-cron-job');``` If possible I would try a fast restart of the database under **Settings -> General -> Restart project (arrow) -> Fast database reboot** Or worst case, a full project restart.

For the function search path thing is basically you likely have no search_path defined. Which means if it runs with Security Definer you've now given permissions for that function to go everywhere and with elevated privileges. Which is is really bad and open to an outside attack. The best solution is to lock it fully down with: **SET search_path = ''** that does mean though that when you build your queries you need to define exactly where they need to look. Pretending users table is stored in the public schema. You can't just do this anymore: **SELECT * from users** You will now need to do this: **SELECT * from public.users**

Pretty sure the RLS policy message means your checks are basically going "everything is approved" which basically isn't much of an RLS policy. They should be limited to exactly what your users need otherwise they could do stuff like delete other's data depending on what policy you setup. When creating a new policy my recommendation is using the Templates on the right hand side to start. Like in this pictures here I only allow authenticated users to view their own data and insert their own data.

I'm guessing you did go over the egress quota very quickly, but the metrics didn't update fast enough. So you delete the old project that would have shown the high usage before you could actually see the results.

I would need to see the code tbh before I could say why. One question and this is more general inquiry so don't take this the wrong way. Is there a particular reason you're using React? Do you have much JS/TS experience?

I mean I can quickly take a look once more at with the latest changes you've made. But that really doesn't teach you how to use react hehe

vite hmr is updating the app on the fly with your changes when you save them

At the very top there's not a drop down of potential other selections is there? My only thought about it disappearing based on what gary is suggesting is that your UIs are stuck on a branch you were working off of and the main project is still there.

The project is paused or pausing? If paused has it been more than 90 days?

My recommendation is retrieve the data, verify its ready, then call setUser once it's in a structure and loaded how you want.

Thus why when you tie into supabase you're having odd situations because your data is just fighting with each other on who should be in state

From a quick review this looks like where data isn't ready yet. You need a loading state to verify everything is ready before displaying. Additionally you have potential race conditions going on here since you have 3 useEffects hooks all trying to update the user state which even if 1 is partially set it can prevent updates from the others.

What I would personally do is watch the database CPU stats and the second you are capping at 99% do this query repeatedly: ```SELECT * FROM pg_stat_activity;``` Watch those query results and look for two things. 1. Is any of the running processes line up with the second the the CPU spiked. 2. While queries will be coming in and failing, keep watching for lets say 3-5 minutes. Are any processes still going after several minutes. Just something that is not ending? Because when I see that graph it feels more like something is getting stuck doing something. Even though it wasn't before hand and an instance increase isn't helping. If a process or something is just eating your instance alive for w/e reason, hopefully this points it out.

Did you grant access to it? ```GRANT EXECUTE ON FUNCTION public.search_users_search() TO "authenticated";``` or anon if its for anon

1. Are you sure you installed the RPC into the public schema? If you go to the dashboard and then Database -> Functions -> Schema (public) is it listed there? 2. On your node.js query are you specifying the public schema? ```await supabaseClient .schema('public') .rpc('your_rpc_here', {```

Let's backup for a second based on that last message. What's your database stats right now? Under observability -> database. How is the memory, CPU, and disk doing?

If you go under Observability -> Query Performance. Then click "reset report" next to export (maybe export first or review first if you desire to see the past) After that do your queries again and see if the DB is actually reporting high times. Additionally on the Project Overview page, are there any issues that need attention in the performance section?

Since that first picture shows to be a select query for all with ordering. If you do that exact same query in the SQL Editor does it also take a very long time? Also how many rows you trying to return?