we might opt for that approach depending on the complexity of handling file chunks
we might opt for that approach depending on the complexity of handling file chunks
Dhonestly if r2 had events, it can be much more powerful and a real competitor to s3
KThats rather simple. You split up the file in chunks on the client and send it one by one
KI go long ways to not use S3 in any projects. 
Dright now, the main value proposition is pricing 
DI'm on GCP so it would be GCS for me haha
Was long as you pay for your usage nope.
Windeed, i said usage not egress :p
WWthe prices listed on that page yes
EGood question
Wtell 'em all to move to R2 
Ř
PHi sorry to interrupt but any fix for this?
KNope - you would need to read the file and then reupload it with the new content
Ucreate a new object and then list and read all 
EEven AWS S3 doesnt support it, there's no API mechanisn for append afaik
RHello. I've noticed an unexpected behaviour when creating R2 API Keys. Seems that API keys are bounded to each profile (user account) but not to the organization account?
This is an issue in case a developer that has created API keys leaves the company then we will lose immediate access to R2. Does someone know how to create an account wide API key?
This is an issue in case a developer that has created API keys leaves the company then we will lose immediate access to R2. Does someone know how to create an account wide API key?
KThat's how all API tokens work in Cloudflare, they're all associated to the user that created them.
KYou get 10 GB-month of free allowance every month. If you store 10 GB for 30 days, that is your free allowance.
KIf you stored 20 GB for 30 days, you'd only be billed for 10 GB-month.
CIs there a way to check if an object in an r2 bucket exist before making the read request
CI’m assuming no
KYou can do
HEADGETBoth will go to the same place and return the same stuff.
CYeah that’s what I thought 
trying to come up with a solution to tell users what objects are available without creating a separate database
trying to come up with a solution to tell users what objects are available without creating a separate database
CYeah the issue is that it incurs a read unit each time a nonexistent object is requested where as if the object does exist it is most likely already cached in cdn and won’t incur read
K404's are cached on the CDN for custom domains
COh really?
CI didn’t know that
CIdk if that will still solve my problem tho. I need the users to see which objects are available and which objects aren’t from the time the visit the website. Right now I’m just using a mongoDB array to store a value indicating whether the object exists. But with egress fees on mongoDB it is so expensive to send each user this data
CAnyone happen to know a different approach? A way to tell users what data is available and what data isn’t
CWhat does that mean
CDo you mean like request multiple objects at once?
COh also, does anyone know if there is a way to update the cache once an item on the r2 has been updated?
CLike say I upload test.json to r2
CAnd that gets cached on read
CThen I make modifications to test.json and upload the new version but it’s still called test.json
CIs there a way to like flag the new version so that the cdn disposes of the old one and request the new one
EThe cache is the same as the regular cdn cache and no unless you send a purge request over api
COk
CCan you explain?
EThe Workers API for R2 is able to list the objects that are in a bucket based on a defined directory/prefix
