Slower

K

kianOP•4/13/24, 9:50 AM

KV does not perform like an in-memory datastore, such as Redis. Accessing KV values, even when locally cached, has significantly more latency than reading a value from memory within a Worker script

Kkian Slower

B

BrianJM•4/13/24, 12:05 PM

What's faster... KV or DO?

K

kianOP•4/13/24, 12:08 PM

Usually KV, but it depends.

K

kianOP•4/13/24, 12:09 PM

Are you creating the DO in Region A and accessing it in Region A? DO will probably be faster

K

kianOP•4/13/24, 12:09 PM

Accessing it globally? KV will be faster due to local caching + central stores in 2 regions

B

bun•4/13/24, 1:02 PM

kv is very expensive

B

bun•4/13/24, 1:02 PM

for no reason

J

juanferreras•4/16/24, 10:15 PM

is using KV metadata simply incompatible with the dashboard? I'm tempted to use metadata for

list()

list()

calls to be self-sufficient, but I worry we'll create a new entry or update an existing one using the Cloudflare Dashboard and we'll have metadata !== value

(or, are there any easy ways to view/edit metadata within the dashboard?)

J

James•4/16/24, 11:09 PM

There's no way to view/edit KV metadata within the dashboard today, nope. I'm not entirely sure honestly what happens if you edit a key's value in the dashboard that already has metadata - that's probably worth testing.

BBrianJM What's faster... KV or DO?

U

Unsmart•4/16/24, 11:27 PM

I'm a bit late to the party but I would personally recommend you always use DO over KV personally. Especially when you can create a DO per user as DOs are faster than KV when close to the user. Only time I use KV is when I am too lazy to bring up a wrangler project and just make a small worker via dash lol.

U

Unsmart•4/16/24, 11:28 PM

Only time KV is really more useful than DOs are when you have larger values but then I would just recommend R2

森

森@VerseLink (RW)•4/17/24, 2:48 AM

would you say DO or KV would be better for reference tokens? (so that we can check up to see if a token is valid or not)

W

W7T2A•4/17/24, 12:49 PM

Is there any option for exporting all data as a backup / or a builtin backup functionality for a kv store? Or is the only way to do this by creating a script via API (for now)

Original message was deleted

C

Chaika•4/18/24, 5:29 PM

you can get storage by namespace analytics via api: https://developers.cloudflare.com/api/operations/workers-kv-stored-data-analytics-query-stored-data-analytics

Cloudflare API Documentation

Interact with Cloudflare's products and services via the Cloudflare API

C

Chaika•4/18/24, 5:30 PM

can't do requests by namespace though

S

sdan•4/19/24, 3:12 AM

anyone know if its possible to use KV within a nextjs app or would i need to make a worker to interface with KV and expose an api from that worker. wondering if the Binding works for nextjs

R

Rubi•4/19/24, 3:37 AM

should I use KV or Durable Object for a quick session check?

R

Rubi•4/19/24, 5:57 AM

let me stay on D1

RRubi should I use KV or Durable Object for a quick session check?

B

bun•4/19/24, 9:39 AM

DO's are cheaper and faster (for a single region)

Bbun DO's are cheaper and faster (for a single region)

B

bun•4/19/24, 9:39 AM

but ye, D1 is fine, i have it on d1 as well

Bbun DO's are cheaper and faster (for a single region)

R

Rubi•4/19/24, 9:54 AM

do ia faster than KV?

RRubi do ia faster than KV?

H

Hello, I’m Allie!•4/19/24, 9:55 AM

If the DO is near you, then yes

R

Rubi•4/19/24, 9:56 AM

alright, gonna using DO

RRubi alright, gonna using DO

B

bun•4/19/24, 10:12 AM

offtopic but your pfp goes insanely hard

M

mau•4/20/24, 12:30 PM

I am implementing cache purging function but hitting a roadblock

for example

http://localhost:8787/api/addons?where[id][equals]=66221339a9cb9c462a245943 caches data from http://localhost:3000/api/addons?where[id][equals]=66221339a9cb9c462a245943

and i am using cloudflare kv for caching. I am generating the kv key using below hash function

export async function getHashString(input) {
  const encoder = new TextEncoder();
  const data = encoder.encode(input);
  const hashBuffer = await crypto.subtle.digest('SHA-256', data);
  return bufferToHex(hashBuffer);
}

function bufferToHex(buffer) {
  const view = new DataView(buffer);
  let hex = '';
  for (let i = 0; i < view.byteLength; i += 1) {
    const byte = view.getUint8(i).toString(16);
    hex += byte.padStart(2, '0');
  }
  return hex;
}

export async function getHashString(input) {
  const encoder = new TextEncoder();
  const data = encoder.encode(input);
  const hashBuffer = await crypto.subtle.digest('SHA-256', data);
  return bufferToHex(hashBuffer);
}

function bufferToHex(buffer) {
  const view = new DataView(buffer);
  let hex = '';
  for (let i = 0; i < view.byteLength; i += 1) {
    const byte = view.getUint8(i).toString(16);
    hex += byte.padStart(2, '0');
  }
  return hex;
}

export async function getHashString(input) {
  const encoder = new TextEncoder();
  const data = encoder.encode(input);
  const hashBuffer = await crypto.subtle.digest('SHA-256', data);
  return bufferToHex(hashBuffer);
}

function bufferToHex(buffer) {
  const view = new DataView(buffer);
  let hex = '';
  for (let i = 0; i < view.byteLength; i += 1) {
    const byte = view.getUint8(i).toString(16);
    hex += byte.padStart(2, '0');
  }
  return hex;
}

export async function getHashString(input) {
  const encoder = new TextEncoder();
  const data = encoder.encode(input);
  const hashBuffer = await crypto.subtle.digest('SHA-256', data);
  return bufferToHex(hashBuffer);
}

function bufferToHex(buffer) {
  const view = new DataView(buffer);
  let hex = '';
  for (let i = 0; i < view.byteLength; i += 1) {
    const byte = view.getUint8(i).toString(16);
    hex += byte.padStart(2, '0');
  }
  return hex;
}

Can someone suggest a method for clearing the cache in the scenario described? I know how to handle cache invalidation for individual items as it's straightforward, but I'm unsure how to effectively clear the cache for queries that involve collection filtering, such as those with multiple conditions. What strategies could work for this more complex case?

Ggrimpa Is there an easy way to know how many values I have stored in KV (in the dashboa...

M

Matt Silverlock•4/22/24, 5:11 PM

Not yet (in the dash) — we're working on a new metrics dash and more metrics for KV.

But you can query via API: https://developers.cloudflare.com/api/operations/workers-kv-stored-data-analytics-query-stored-data-analytics

    "max": {
      "storedBytes": 32,
      "storedKeys": 4
    },

    "max": {
      "storedBytes": 32,
      "storedKeys": 4
    },

    "max": {
      "storedBytes": 32,
      "storedKeys": 4
    },

    "max": {
      "storedBytes": 32,
      "storedKeys": 4
    },

Cloudflare API Documentation

Interact with Cloudflare's products and services via the Cloudflare API

V

Varna•4/25/24, 1:06 PM

any recommendations for auto generated keys for KV? I just want to store two weeks of reports in sequential order (few hundred records tops)

VVarna any recommendations for auto generated keys for KV? I just want to store two wee...

H

Hello, I’m Allie!•4/25/24, 1:09 PM

Date.now()?

HHello, I’m Allie!Date.now()?

V

Varna•4/25/24, 1:11 PM

thought about that too

but maybe it would be better to use smth like

crypto.randomUUID()

crypto.randomUUID()

crypto.randomUUID()

crypto.randomUUID(), or maybe I'm overthinking this

V

Varna•4/25/24, 3:13 PM

I wanted to test if

expirationTtl

expirationTtl

expirationTtl

expirationTtl works by setting it's value to 70 (seconds).
It did set it to

expiration: 1716649322

expiration: 1716649322

expiration: 1716649322

expiration: 1716649322 on

KV.list()

KV.list()

KV.list()

KV.list() response which is 18:02 (8 minutes ago).
Is it purged at batches on some intervals, or do I need to do that manually?
Or is this normal if I had a long