I do want to get to make list operations more like R2, but I'm not sure we'll be able to do that thi
I do want to get to make list operations more like R2, but I'm not sure we'll be able to do that this year.
BStrong candidate for 2024 though.
BI know that's not a great answer though in the short term.
KAh could you give me an example of the syntax that would be used to accomplish my goal? BTW I am building a new company on cloudflare and you guys are going to destroy every cloud provider this decade
SI’m just waiting for CF Auth and CF Mail 
ACf mail is mailchannels?
Eand #email-routing for the other direction
BThere’s no way to do it from your worker today. I would imagine we’d unify it similar with how list works on R2
CYou are correct. The node library is very out of date. The last commit was a year ago and the last tag two years ago
BWow I didn't even realize that was a thing. I'll raise it internally to see what the plans are for that.
CYeah. The go api is the most update if you know go
CIIRC, there is a plan to auto generate clients once the shift to openapi is done. The node, python, and PHP libraries don’t receive a lot of love otherwise
Mcan bulk KV write operations occur simultaneously? i'm writing around 128K items to a KV store in batches of 2000 over a period of about a minute and i'm noticing that some values aren't getting written.
Mat the moment i'm doing some further debugging to make sure that it's not me doing something dumb, but i figured i'd check here as well.
Mmystery solved: 429 Too Many Requests
Mis rate limiting per IP or per API key?
JAccount ID in general, I believe
J429s are pretty common when doing bulk stuff with KV
CYeah API rate limit is per account
M
tyMis rate limit info posted anywhere? googling for anything cloudflare-related + 429 gives me a whole lot of useless results
ETheres this which says 1200 req per 5 minutes, but I dont think thats the one youre running into? https://developers.cloudflare.com/fundamentals/api/reference/limits/
Myeah i don’t think that’s the one i’m getting
CI thought it would be. Wrangler just uses the normal API when uploading in bulk
Minnteresting… maybe it is? i’m definitely not sending 1200 requests in five minutes tho. 64 calls containing 2000 items each.
JIt's not uncommon for KV to throw 429s with bulk data unfortunately, even outside of the API req limits 
Moh dear. good to know, thank you. i’ll add some extra handling on my end.
B64 concurrent calls with 2k items each? Yeah, that’s >100k writes/s if my math is right. That’s too high right now. I wouldn’t recommend going higher than 10k/s right now. KV2 will give us more breathing room. I do want to get to much higher later this year but it’ll depend on internal arch stuff.
Mahhhhhhh that makes sense, i wasn’t thinking of writes, just requests
OHey folks, I am seeing something (strange?) - KV has much higher latency than R2, is that normal?
To back it up, here is what I am looking at.
For static JS, we are using workers for platforms and we use kv-asset-handler
For fonts, we use fetch from R2
Fonts are waiting for response under 50ms consistently
Scripts are waiting for response from 150ms to 500ms
We tried multiple regions, I am personally Berlin.
To back it up, here is what I am looking at.
For static JS, we are using workers for platforms and we use kv-asset-handler
For fonts, we use fetch from R2
Fonts are waiting for response under 50ms consistently
Scripts are waiting for response from 150ms to 500ms
We tried multiple regions, I am personally Berlin.
BKV should not have higher latency than R2 generally. R2 can have better latency if you happen to be in the same region you placed the bucket (which is generally the default if the machine you created the bucket is the same machine making these requests). Even still, for an asset you fetch repeatedly KV should still be faster.
BIf you're changing the UUID of the script frequently, that would trigger a cold read every time. So this is very dependent on your test experiment environment.
OWe are not changing the UUID of the script during the tests, so its really static. https://webstudio.is
BOk. Let me take a look
BSo just so I'm clear, for example https://webstudio.is/build/entry.client-QHBMJZIF.js is taking 137ms in your screenshot and that's served from KV?
BSo looking at our backend, it looks like there's no KV asset for that file. It looks like you might be trying to put cache in front of KV but then forgetting to update KV on a miss?
IIts worker for platform, it has following code.
KV is created by using
section in toml file
and exists and filled
KV is created by using
section in toml file
and exists and filled
4453161ad26e32a617c6c6d4b837c9d9/workers/kv/namespaces/989796259c334124b2e868bd9a0c3849This library https://github.com/cloudflare/kv-asset-handler std code is used for static assets hosting
Itimings are very unstable and sometimes can be very good
IBut next reload
I
IFont timing (r2 with cache rule) is independent
BThat library seems to have a lot of magic and I've never looked at it before
OThe weird thing is that R2 has consistently good timing. Should we put everything in R2 then? 
BNo. I'd recommend using KV properly
ICan it be because of worker is blocked by some heavy script, like first request is coming and cause some heavy job on worker - really 100-200ms computation, and consequent requests are waiting for worker finish that task?
BNo
