Worker responses aren't cached, what are you referring to?

Zzegevlier Worker responses aren't cached, what are you referring to?

SSDD•12/20/23, 10:30 AM

I have a worker deployed.. which executes graphql requests, it fetches data from other website which takes 1-3 secs.

1st request takes around the expected time but after that it's always ~300ms response time

SSDD•12/20/23, 10:32 AM

and I do not have any type of caching in the code that I deployed

zegevlierOP•12/20/23, 10:44 AM

300ms seems very slow if it was actually caching the worker responses. Perhaps the GraphQL server your code is contacting is caching it?

mukhtharcm•12/20/23, 10:44 AM

Quick question #urgent

the daily limit of 100k request is a hard limit?

will the worker fail after that?

zegevlierOP•12/20/23, 10:44 AM

Yup, hard limit.

zegevlierOP•12/20/23, 10:45 AM

https://developers.cloudflare.com/workers/platform/limits/#daily-request

Zzegevlier Yup, hard limit.

mukhtharcm•12/20/23, 10:46 AM

Okk

Zzegevlier 300ms seems *very* slow if it was actually caching the worker responses. Perhaps...

SSDD•12/20/23, 12:28 PM

I don't code is caching.. ig I can put manual delay for each response to see if cf is caching

liam | helius.dev•12/20/23, 7:45 PM

What would cause a load balancer using proximity steering to send requests to pools in different locations (besides if a pool is unhealthy)? Shouldn't it be static?

Sam•12/20/23, 8:13 PM

Is it possible to combine a cronjob, queue and web api (hono) project in one folder?

Eg:

export default {
  fetch()
  scheduled()
  queue()
}

export default {
  fetch()
  scheduled()
  queue()
}

SSam Is it possible to combine a cronjob, queue and web api (hono) project in one fol...

Hello, I’m Allie!•12/20/23, 8:21 PM

Yes

Unsmart•12/20/23, 11:43 PM

Pretty sure you dont want to use clone for that.

Should just be:

new Response(response.body, { headers: {} })

new Response(response.body, { headers: {} })

https://developers.cloudflare.com/workers/examples/modify-response/

Skyy•12/21/23, 1:54 AM

Is there anyway to run a CRON trigger every 30 seconds? Or 2 with 1 being 30 seconds delayed?

SSkyy Is there anyway to run a CRON trigger every 30 seconds? Or 2 with 1 being 30 sec...

0x0dd154•12/21/23, 1:19 PM

await a promise in one job that resolves after 30s?

I AM UNDER ATTACK•12/21/23, 1:46 PM

hello, from 2 weeks i am facing high traffic load, which is comming from ddos attact, i have already enabled the proxy and i am under attack mode enabled, but still facing ddos attack on my server, my use goes very high, please help me to get out from this

00x0dd154 await a promise in one job that resolves after 30s?

Skyy•12/21/23, 1:48 PM

Worker Cron jobs have a max time of 30 seconds, So that wouldn't work?

I AM UNDER ATTACK•12/21/23, 1:51 PM

SSkyy Worker Cron jobs have a max time of 30 seconds, So that wouldn't work?

0x0dd154•12/21/23, 2:03 PM

Good point. This is probably a good use case for DO alarms, but if that's not already part of your toolbelt would be introducing more moving parts.

00x0dd154 Good point. This is probably a good use case for DO alarms, but if that's not al...

Skyy•12/21/23, 2:42 PM

Currently I'm not using any durable objects, So yeah, It would be implementing extra items to the workflow

SSkyy Is there anyway to run a CRON trigger every 30 seconds? Or 2 with 1 being 30 sec...

Chaika•12/21/23, 2:45 PM

DO Alarms would probably be easiest.
You can't use crons like that sadly, they're just not accurate on the seconds and as they move between being scheduled in one cloudflare location/colo they can shift by 10-30s

CChaika DO Alarms would probably be easiest. You can't use crons like that sadly, they'r...

Skyy•12/21/23, 3:04 PM

Ah, I was thinking it would have an offset at most like 10s, As I didn't really need precision timing. Okay, I'll look into DO

Skyy•12/21/23, 3:05 PM

Thank you guys!

jkomyno•12/21/23, 4:07 PM

Hi! Let's say I have a large-ish Wasm file (> 2MB) which I want to import in a JS cloudflare worker.
Such Wasm file goes down to a few hundred KB (< 1MB) after applying zstdzstd compression locally.
Is there a way of telling my CF worker to apply zstdzstd decompression when uploading my project's code to the hosted v8 isolate?
I've been looking at https://developers.cloudflare.com/workers/wrangler/custom-builds/ but I couldn't find a clear answer

Custom builds · Cloudflare Workers docs

Customize how your code is compiled, before being processed by Wrangler.

kian•12/21/23, 4:09 PM

Nope, Workers uses gzip

Kkian Nope, Workers uses gzip

jkomyno•12/21/23, 4:17 PM

I see, and alternatively, if I used a JS-only zstdzstd implementation, would I be able to apply that to a Uint8ArrayUint8Array representation of a Wasm binary embedded in a JavaScript constant, and call WebAssembly.instantiate()WebAssembly.instantiate() from it?

samy•12/21/23, 4:17 PM

When I activate 1.1.1.1 and warp it doesn't work

Jjkomyno I see, and alternatively, if I used a JS-only `zstd` implementation, would I be ...

kian•12/21/23, 4:25 PM

Nope, Workers doesn’t allow code generation which includes eval/new Function/WebAssembly.instantiate(notAModule)

Village (Josef)•12/21/23, 4:29 PM

Is there any way to enable new Function()new Function() on Workers? It turns out a library I'm using relies on it for it's core functionality and I'd rather not reimplement it all

Village (Josef)•12/21/23, 4:32 PM

The feeling when the question above me was the same thing and I didn't notice dead

VVillage (Josef)Is there any way to enable `new Function()` on Workers? It turns out a library I...

Unsmart•12/21/23, 4:33 PM

Nope its turned off for everyone

Village (Josef)•12/21/23, 4:33 PM

Now I have to go reimplement a thing

Peps•12/21/23, 4:43 PM

The docs for the new sendBeacon api appears to have issues

Screenshot_2023-12-21-17-43-03-204_com.android.chrome.jpg

space77along•12/21/23, 4:49 PM

I can see the WebSocket support is gradually improved over the time, as well as mqtt support, I wonder how long will WebSocket connections remain open once connected from client? It is not immediately clear to me by reading: https://developers.cloudflare.com/workers/examples/websockets/

Using the WebSockets API · Cloudflare Workers docs

WebSockets allow you to communicate in real time with your Cloudflare Workers serverless functions. In this guide, you will learn the basics of …

Unsmart•12/21/23, 4:50 PM

It depends how long it takes for you to reach the CPU limits

space77along•12/21/23, 4:50 PM

OK? How do I calculate that?

space77along•12/21/23, 4:51 PM

I'm trying to replace a node-red implementation that receive from mqtt and send to WebSocket connected to that topic at any moment, basically an adapter. No storage is required.

Unsmart•12/21/23, 4:52 PM

There's not really any good way to know your CPU usage unless you are enterprise unfortunately.

space77along•12/21/23, 4:52 PM

We are enterprise, is this something we need to purchase as addon?

Unsmart•12/21/23, 4:55 PM

You can use http logpush and enable the workers cpu time field, that will get you per request cpu time. Havent tested it myself but websockets specifically should show up after it is closed with the full cpu time it took during the entire time it was connected. So you can open up a websocket throw some sample requests you would do and close it. Take note of the cpu used and the duration and use that to see max length of connection possible by the cpu limit you have (30s max and default)

space77along•12/21/23, 4:57 PM

OK thanks, should be cheaper than IBM WebSphere anyway haha.

ArjunBroEpic•12/21/23, 5:53 PM

I keep getting ExceededCpu 503 on a route using Hono and lucia-auth to login a user with D1
I have tried to optimize
It randomly works and stops working

ArjunBroEpic•12/21/23, 6:11 PM

nvm skill issue optimized even more

thiccvitalik•12/21/23, 6:57 PM

Hi! I asked this in queue-beta, hope it's ok to ask it here too: When I have a worker that is a queue consumer with a max_batch_size of 5, does this imply that the processing of those 5 messages will share the 128MB memory allocated to the worker? And if I set the max_batch_size to 1, does it ensure that the processing of each message will exclusively have 128MB of memory?

Orlando Becerra•12/21/23, 7:13 PM

Hi! I have created a worker using the web dashboard, now several developers need to be able to make changes to this project locally, we followed the steps to install wrangler, but we can't find a way to fetch/download this project/worker to make changes to it locally, does anyone know how to do that? thank you

OOrlando Becerra Hi! I have created a worker using the web dashboard, now several developers need...

Adam•12/21/23, 7:48 PM

You can use wrangler init --from-dash <WORKER_NAME>wrangler init --from-dash <WORKER_NAME>

AAdam You can use `wrangler init --from-dash <WORKER_NAME>`

Orlando Becerra•12/21/23, 7:49 PM

thank you, checking it

AArjunBroEpic nvm skill issue optimized even more

ArjunBroEpic•12/21/23, 8:31 PM

nvm it's still there
I cannot optimize it any more and it takes around 100ms

Worker responses aren't cached, what are you referring to?

Similar Threads

Similar Threads

Similar Threads