I only have anecdontal evidence of people showing pro/biz helps dtag/india routing. I will add some after this because I'm curious, but even when I do the data I have is all from datacenter connections, so not 100% representative of real isp connections
How is workers cpu time calculated? Is it possible that the cpu time calculation strategy was adjusted in October, and I realized that my bill does not correspond to it?
Hi @Walshy, I got question regarding the workers ai and vectorize, I have an app running Resnet50 to vectorize image, I'm planning to switch over to cloudflare since you guys support Resnet50, however, I saw the limitiations of the Vectorize it says it has maximum dimensions per vector 1536 dimensions, which is not high enough for my app since it generate 2048 dimensions using Resnet50, any idea how can I reduce the dimensions the is generated from Resnet50?
I'm really new to this field, so any help would be awesome, thanks!
I am using this approach, modifying the date query. I used this general method to look up the cpu time for my November bill and it matched, but not for October.
The datetime filter in the second image is in the middle of a billing period. Do something like Oct1 to Oct30 or something to see that month. If you have not a whole month of usage, it will only be the portion of the month you were using
You seem to have it the wrong way around? The 6 524 number is what you're being billed for while the 61 442 number is from GraphQL, so if anything, you're being billed 10x less than what GraphQL shows?
If the mismatch is present in September there's no reason to think it won't also be in October/November. Nothing has changed regarding pricing of Workers so if there is a mismatch I'd expect it's either an issue in the billing system (possible, billing system has been undergoing some changes recently) or an issue with the query.
What I'm wondering is if the actual billing data increased by 10x on your invoice, or if it's just a mismatch between GraphQL and the invoice? From the screenshot you sent it looks like it didn't for September, did you observe this for October or November?
I'd be interested to see if you could figure out an average CPU MS per invocation in September then October to compare, see if each instance is using roughly 10x or if theres some small amount of instances hogging time