similar to the 128mb limit for stateless workers

Ah great.

Interesting. But doesn't invoking KV , even when "hitting the cache" (as this is transparent to the Developer) always incur costs, while using cache API directly doesn't?

Yes, that's true

This is what @HardAtWork 's BetterKV wrapper does under the hood: check out

flareutils

on npm docs here: https://flareutils.pages.dev/classes/BetterKV

I am trying to create an API in cloudflare workers. I want to bill my API based on how many requests the subscriber uses. For example, Person A signs up for a plan that allows 100 requests per day. How would I prevent Person A from going above that limit? Essentially, how do I keep track of a simple +1 increment without really expensive KV writes? External Databases also aren't an option because that eliminates the entire reason I'm using edge functions.

You need Durable objects.

already got my answer in #779390076219686943 but thanks!

I ended up using analytics engine

Okay. How’d you implement that ?

not implemented yet

but it seems fairly simple

Cause I’m trying to do the same thing. But I don’t know about the Analytics engine approach

read analytics engine docs & https://discord.com/channels/595317990191398933/779390076219686943/1058881949147607140

Well with hard limits it's going to be impossible to use AE because of rate limits on the read end.

Which is why I suggested DO > AE

with "hard limits"

you can allow a little bit of wiggle room

at least in my case

Well even a soft limit because you still need to read the total number of requests they've done and you can't query AE often.

what's the read limit speed on the AE

in fact, you don't even need to use WHERE

1200 reads every 5 minutes

you can just use an SQL search that finds all things that are above 500 requests

no need to use a ton of WHERE requests

every second, just do this

correct me if I'm wrong ofc

Having many users you will easily reach that 1200 count. Especially since not all data might fit on one page

"reads", as in?