You can list by prefix and then you can randomize the results on the client side fairly easily?
You can list by prefix and then you can randomize the results on the client side fairly easily?
MWorkers KV is a global, low-latency, key-value data store. It stores data in a small number of centralized data centers, then caches that data in …
MSpecifically, just choose a random integer between 0 and
length of your list -> then access that array item directly. 
ROk, it makes sense.
Btw, I've checked the
https://github.com/helloimalastair/flareutils/blob/0aa471d/src/BetterKV.ts#L295
But I see it uses this URL
Do you know what this URL is about? Am I supposed to get the same results as using the KV alone? Thx
Btw, I've checked the
flareutils library and I see there's also a list method combining cache and kv. https://github.com/helloimalastair/flareutils/blob/0aa471d/src/BetterKV.ts#L295
But I see it uses this URL
cacheKey = new URL("https://list.better.kv"); Do you know what this URL is about? Am I supposed to get the same results as using the KV alone? Thx
GitHub
Contribute to helloimalastair/FlareUtils development by creating an account on GitHub.
HThat url is just there to uniquely identify the list operation in the cache. You can call it whatever you want, as long as it is unique
JYou won't be able to run something like
axios on Workers, which is sounds like you're trying to doJYou'll need to use
fetchWorkers KV is a global, low-latency, key-value data store. It stores data in a small number of centralized data centers, then caches that data in …
Jaxios uses Node's http module, which doesn't exist in Workers. Workers run JavaScript via V8, but not Node.jsJI'm not sure if there's any specific literature on KV + TS, but it's all built into
@cloudflare/workers-types. There's a KVNamespace type which you can assign to env.KV_BINDING or similar, which should work well
Wfetch
Uits not a default of the runtime
WAxios a few years ago was great, these days
fetchCan you provide a reproduction? A link to your repo or something and the commands you’re running would be good
JSure. It’d probably make sense to move to #workers-help to condense this all 
Ahi, it seems there is nothing returned from kv put call. how do i know put is successful ?
Whttps://developers.cloudflare.com/workers/runtime-apis/kv/#parameters
Keys have a maximum length of 512 bytes.
EThere's a cost-consideration to the customer. Delete operations are capped to 1k/day for free customers, and 1mm/month + 5$/million for paid accounts. Deleting a KV namespace is ~free~.
GHi !
The documentation ( https://developers.cloudflare.com/workers/learning/how-kv-works/ ) say « Frequent reads from the same location return the cached value without reading from a central data store, resulting in faster response times. »
There is a more precise answer of what is considered « frequent reads » please?
To known if using KV for us is big mistake for performances or not
The documentation ( https://developers.cloudflare.com/workers/learning/how-kv-works/ ) say « Frequent reads from the same location return the cached value without reading from a central data store, resulting in faster response times. »
There is a more precise answer of what is considered « frequent reads » please?
To known if using KV for us is big mistake for performances or not
Workers KV is a global, low-latency, key-value data store. It stores data in a small number of centralized data centers, then caches that data in …
HUnless you mess with the CacheTTL, reads will be cached at most 60 seconds
HDerp
DIs there anything cheaper than
Obviously, our current solution is not ideal as it requires a re-deploy to change environment variables. Any better solutions that are a bit cheaper than
kv for super simple storage on Workers? Basically, we have a boolean that we need to toggle occasionally. The server is doing 10s-100s of millions of lightweight requests per day so KV is surprisingly expensive for this use case. We end up using an environment variable and flipping that variable every time we want to change the value. Obviously, our current solution is not ideal as it requires a re-deploy to change environment variables. Any better solutions that are a bit cheaper than
kv?DPerhaps we could put a cache layer in front of
kv and then we could invalidate every minute or so? Would that be a good solution?DBut you could automate the task of changing it quite easilyWe need to re-deploy every time we change an
env right? We can't do it on the fly?DNice! Cache is basically free right? Or, to be more correct, we just pay for the increased duration of the function.
DCool,
kv with a cache ttl seems like the right solution. Any thing I'm missing?DPerfect. Thanks @skye_31, appreciate your help. 
HIndefinitely
JPerhaps D1? I know it's in alpha. If not that, then maybe PlanetScale? Their "Scaler" plan would give you 100 billion row reads/mo.
HA Durable Object maybe?
