was hoping there was something simpler

Current solution is a WIP. I store an array of corresponding r2 file ids in my DB and then check that to see I should generate another presigned URL for the user. The issue is that if they're under the max, they could hit that endpoint multiple times to generate multiple presigned URLs and then upload over the limit from there. It only works for if someone is already at the max 😬 .

You would just not allow generating a new presigned url if they are at the max, or instead of using presigned urls use your origin/a worker for uploads

gotcha thanks. I will give that a shot!

presigned urls are a bit interesting because you would want to create a background job to see if the file was actually uploaded once it expires and if not then allow 1 more

I suddenly got

UnauthorizedUnauthorized

when accessing my R2 with presigned url 😐

the same app run fine yesterday, no code changes, no recent deploys

Turns out it was my R2 api token expired, although I set it end date to null when I first create it 🤔

We just got hit with a Section 2.8 violation for content hosted on R2. We're storing large ML datasets as well as images and videos. My understanding from talking to the R2 team previously is that heavy users of R2 must proxy requests through a custom domain through Cloudflare (versus the r2.cloudflarestorage.com endpoint). However, it seems like it's triggering this violation? My understanding is that R2 is covered under the supplemental terms so I was surprised to receive this notice regardless. I know several teams storing many TiB of data on R2 such as parquet, so would like to clarify this.

> My understanding is that R2 is covered under the supplemental terms It is > However, it seems like it's triggering this violation? It will - the detection isn't very smart and doesn't currently pick up that you're using R2 which is an allowed product to serve non-HTML content. You'll need to open a support ticket and then they'll be able to get it lifted for you.

If you post the ticket number here, I can get it escalated so someone from CSUP can take a look & send it to the right team

Got it thanks. It would be great to have clarification in the docs on how to pre-clear R2 with support since it brought down our website. We already notified our userbase that Cloudflare R2 is the cause of the outage.

The case number is 2749735, thanks.

I've raised it with CSUP - I have no insight into how long the process of getting it lifted takes, but expect updates to go through that ticket.

thank you. it's a website to host some of our research (we're UC Berkeley PhD students) and it got a massive hug of death from twitter yesterday.

How can I track R2 object downloading?

Like, file A got streaming 15MB, file B got streaming 20MB, ...

Is the automatic detection of non-HTML ever going to be updated to take into account R2? R2 is literally a product for serving non-HTML requests. I've had a few of my users get hit with the same issue. It's starting to leave a bad taste in user's mouths... "Hey, move all your images and attachments to R2! You are in violation of Cloudflare terms for moving images and attachments to R2, so your site is broken until you open a ticket." Now I have users that are thinking about moving to R2 deciding they don't want to because they see the support requests from users that did and got hit with (temporary) violation suspensions.

Hopefully. This is an actively monitored thread right now so if you're on Twitter, I'd recommend sharing your experience there: https://twitter.com/_parasj/status/1641470460440481793

We're just shy from a year after R2's open beta so it is a bit disappointing it isn't fixed yet but yes, they are planning to stop it from happening. Like James said, that Twitter thread has been raised internally.

Well it's not my personal experience. I have a few thousand Cloudflare accounts using the integration I built for their sites, so I more deal with those users complaining to me. But at the end of the day, I don't know what zones they are or anything like that (all done via API tokens the user enters, so I don't have access to their accounts). I'll point my users that it's happened to there and maybe they will share.🤷

"I have thousands of users using an integration I built that regularly run into this issue with legit usage" would be great feedback to share I'd imagine, even without specific examples 🙂

I imagine the more people that report this issue, the higher the chances of something being fixed.

I guess it wouldn't hurt to toss that in... just seems weird that I wouldn't be able to give ticket numbers or even the domains it happening to. I don't ask users what their sites are or anything, but I suppose I could if needed when they mention it happened to them

Hello everyone, Is anyone able to tell me if I set a

if-none-match

header it counts towards the PutObject usage if it already exists in the bucket? Thanks

anyone else get this when trying to cash in a presigned url

Access Key refers to the API token access key, right?

is R2 integrated with AWS somehow?

I've never made any sort of amazon account