TBD

to be decided?

Yes, to be decided..

I feel like it wouldn't make sense to, because we could just cache it using the worker cache api for free

i mean, workers aren't free

True, but lots of people will be accessing R2 through the worker bindings

I sure hope there's some easy way to access and control caching without having to rely on a worker that does nothing but 10 lines of code and still getting charged.

Without any number to support it, but just going by feeling I think a significant portion will be "private upload/public download."

i'd imagine there could be differences in the maximums between the cache api and the R2 "cache"

the worker cache api max object size is only 512MB

R2 has a much much larger maximum object size

so it's possible the R2 cache and the worker cache are very different beasts, i'm not really sure

i mean it's possible your usecases for KV/cache/R2/etc are all for small-ish objects. but i'm just pointing out that people keep multi-GB objects in S3 buckets, and given R2 is analogous to S3, it'll need to be dealing with a different mix of usecases than kv, cache api, etc, have so far, at least when it comes to object size. it could be that CF has generalized their caching architecture such that it will be shared across both worker cache and R2, but i don't think that's really been stated

That makes sense, thanks!

Will it be possible to upload files to R2 that are larger than 512MB? AFAIK that's the largest request size with Workers (and even that's Enterprise) so I'm guessing there would need to be a public (non-Worker) API or similar to accomplish that?

i believe the s3 endpoint would support that

This is regarding the automatic S3 pull so I guess so

Oh yeah, perfect

😄

hey all, some delay in adding the people who got access through discord on Monday. I'll update once they're in!

Awesome, thanks Greg!

Nice

Sounds great!

is that right?

I see the link for R2

and it wasn't there before

looks to be