If I'm worried about a Worker getting spammed (not a strong concern, just a thought), is there a way

If I'm worried about a Worker getting spammed (not a strong concern, just a thought), is there a way to protect against that before it hits the worker to save cost? It's not a huge deal since Workers are inexpensive. But are there options?

Ssirhype If I'm worried about a Worker getting spammed (not a strong concern, just a thou...

Cyb3r-Jak3•3/22/24, 9:42 PM

You can use rate limiting rules. https://developers.cloudflare.com/waf/rate-limiting-rules/

Cloudflare Docs

Rate limiting rules · Cloudflare Web Application Firewall (WAF) docs

Rate limiting rules allow you to define rate limits for requests matching an expression, and the action to perform when those rate limits are reached.

sirhypeOP•3/22/24, 9:45 PM

Ah thank you so much!

SageSphinx63920•3/23/24, 11:54 AM

Anyone experienced this issue with email workers before

Unknown error: permanent error (521): 5.5.2 mail.sageee.xyz Error: bare <LF> received

Unknown error: permanent error (521): 5.5.2 mail.sageee.xyz Error: bare <LF> received

SageSphinx63920•3/23/24, 11:55 AM

Trying to sent a message to my mail address with

mail.sageee.xyz

mail.sageee.xyz

being my mail server but i do not know why it is not working as sending mails to smth like outlook or gmail does work

SageSphinx63920•3/23/24, 12:18 PM

This is my postfix log

SageSphinx63920•3/23/24, 7:35 PM

fixed this

SageSphinx63920•3/23/24, 7:36 PM

But now i have this error

could not obtain a forward upstream: Failed to connect to egress transport: no usable transport

could not obtain a forward upstream: Failed to connect to egress transport: no usable transport

when sending mails to a specific mail provider

SageSphinx63920•3/23/24, 7:36 PM

Can anyone help with this

GoodluckH•3/23/24, 8:24 PM

why would i need a queue if i can invoke wokers millions of times a day? let's say every day i have 10000 active users each needs to get info from 100 websites. so a worker basically just scrape the info.

in this case can't i just do a 10,000 * 100 for loop to call my worker 1 million times?

GGoodluckH why would i need a queue if i can invoke wokers millions of times a day? let's s...

Chaika•3/23/24, 8:53 PM

Queues are for more reasons for that,
to start with, you get 1k subrequests per worker invocation, only 6 at one time, and Workers can only call workers up to a certain depth (I forgot what, exactly). The user may disconnect before all your requests are done.

Queues also are for instances where your backend can't handle it at all at once. They let you trade off latency for throughput.

GoodluckH•3/23/24, 8:54 PM

interesting. how did you get 1k subrequests?

Chaika•3/23/24, 8:55 PM

The limits are on here: https://developers.cloudflare.com/workers/platform/limits/

Cloudflare Docs

Limits · Cloudflare Workers docs

Cloudflare Workers plan and platform limits.

GoodluckH•3/23/24, 8:56 PM

got it! thanks!

SSageSphinx63920 Anyone experienced this issue with email workers before `Unknown error: permanen...

DarkDeviL•3/23/24, 11:20 PM

Was this the one you referred to with "fixed this"?

If not, you need to look around in you chain, all from where you are putting the message together, and to the point where you're seeing that error.

Somewhere in that chain, it appears that you're ending lines in your code just with

\n

\n

(

<LF>

<LF>

), which must be changed to

\r\n

\r\n

(

<CR><LF>

<CR><LF>

Silent•3/24/24, 12:54 AM

Is it possible to return a seperate html file as a response and have it show that?

SSilent Is it possible to return a seperate html file as a response and have it show tha...

Walshy•3/24/24, 1:06 AM

Yep, you can import the file and return it (it'll be uploaded as a module)

E.g.

import indexHtml from './pages/index.html';

export default {
  fetch(req) {
    const { pathname } = new URL(req.url);

    if (pathname === '/') {
      return new Response(indexHtml, { headers: { 'Content-Type': 'text/html' }});
    }

    return new Response('404 Not Found', { status: 404 });
  }
}

import indexHtml from './pages/index.html';

export default {
  fetch(req) {
    const { pathname } = new URL(req.url);

    if (pathname === '/') {
      return new Response(indexHtml, { headers: { 'Content-Type': 'text/html' }});
    }

    return new Response('404 Not Found', { status: 404 });
  }
}

(Wrote on phone so won't be perfect)

DDarkDeviL Was this the one you referred to with "fixed this"? If not, you need to look ar...

SageSphinx63920•3/24/24, 7:26 AM

This is fixed the second problem sadly not

SageSphinx63920•3/24/24, 7:27 AM

I have absolutely no clue what’s wrong because I can not find Thai online somewhere

thanhle•3/24/24, 1:49 PM

i really love workers and really enjoy using it. my question is: is there anyway I deploy a worker but dont go to egde network. I want the workers only deployed to a specific region not global. I try to use smart placement but not help. the reason I want to deploy worker to a specific region is when I benchmark the query Postgresql database by using CloudFlareWorker and AWS Lamba(aws lambda is deployed to only one region), the response time for aws lambda is TWICE FASTER than that of cloudflare workers. Thank you all.

alotabits•3/24/24, 2:07 PM

The Cache API allows interaction with the same cache used for fetch requests via caches.default. Is it possible to get or put the entries in this cache for a particular device type in a manner compatible with cache-by-device-type settings?

alotabits•3/24/24, 4:04 PM

So it’s not a special cache key value? As long as we construct a Request with the correct cf-device-type header when passing to cache get/put/delete it will work correctly?

codebam•3/24/24, 6:41 PM

the example code here doesn't work

https://developers.cloudflare.com/workers-ai/models/llama-2-7b-chat-fp16/

where does Response come from??? response doesn't have .json() it's a

ReadableStream<any> | AiTextGenerationOutput

ReadableStream<any> | AiTextGenerationOutput

Cloudflare Docs

llama-2-7b-chat-fp16 · Cloudflare Workers AI docs

Run AI models in Workers, Pages, or via API.

Ccodebam the example code here doesn't work https://developers.cloudflare.com/workers-ai...

Hello, I’m Allie!•3/24/24, 6:49 PM

Response.json()

Response.json()

belongs to the global

Response

Response

class.

Cloudflare Docs

Response · Cloudflare Workers docs

Interface that represents an HTTP response.

HHello, I’m Allie!`Response.json()` belongs to the global [`Response`](https://developers.cloudfla...

codebam•3/24/24, 7:04 PM

oh okay. does that mean I can't do multiple responses at the same time?

Ccodebam oh okay. does that mean I can't do multiple responses at the same time?

Hello, I’m Allie!•3/24/24, 7:15 PM

Not sure what you mean? You can construct multiple

Response

Response

objects at once, but you shouldn't need to for the tutorial you linked

Tony Leung•3/24/24, 8:25 PM

hi guys, is there anyway to use crypto.DigestStream in local dev environment? I'm using svelte and having this error:

TypeError: crypto.DigestStream is not a constructor

TypeError: crypto.DigestStream is not a constructor

codebam•3/24/24, 8:29 PM

getting this error when building

../../node_modules/@cloudflare/ai/dist/ai/tasks/image-classification.d.ts(2,24): error TS2307: Cannot find module '../ai' or its corresponding type declarations.

../../node_modules/@cloudflare/ai/dist/ai/tasks/image-classification.d.ts(2,24): error TS2307: Cannot find module '../ai' or its corresponding type declarations.

codebam•3/24/24, 8:29 PM

on the latest node

codebam•3/24/24, 8:30 PM

it built successfully 3 weeks ago

codebam•3/24/24, 8:38 PM

it doesn't build on the same commit that built successfully 3 weeks ago

alotabits•3/24/24, 8:39 PM

If we have different versions of wrangler installed as a dependency in multiple npm packages locally, each package representing a separate worker, can these workers still be bound to each other via service bindings by starting wrangler dev in each?

WWalshy Yep, you can import the file and return it (it'll be uploaded as a module) E.g....

Silent•3/24/24, 9:24 PM

i have to add type definitions or it manually or?
did not mean to ping&*

bun•3/24/24, 10:57 PM

@HardlyWorkin' since you can purge cache by url, cant you just make

bKv

bKv

cache indefinitely while purging on

.put

.put

requests?

bun•3/24/24, 10:57 PM

only caveat is that you can only cache stuff by

caches.default

caches.default

Bbun @HardlyWorkin' since you can purge cache by url, cant you just make `bKv` cache ...

Hello, I’m Allie!•3/24/24, 10:58 PM

Maybe, but then you are limited to around 1k writes a minute(or whatever the limit is)

bun•3/24/24, 10:59 PM

also cache api only caches on the datacenter its called from right?

bun•3/24/24, 10:59 PM

like it doesnt replicate to other datacenters?

Bbun also cache api only caches on the datacenter its called from right?

Hello, I’m Allie!•3/24/24, 11:05 PM

Oh, you meant like that

Hello, I’m Allie!•3/24/24, 11:05 PM

You could do that too, but as you pointed out, it wouldn't propagate

bun•3/24/24, 11:06 PM

hm what did you mean then

bun•3/24/24, 11:06 PM

initially what u thought?*

Bbun hm what did you mean then

Hello, I’m Allie!•3/24/24, 11:06 PM

I thought you meant purging the Cache via api.cloudflare.comapi.cloudflare.com

bun•3/24/24, 11:06 PM

yes, but the cache set from cache api

Bbun yes, but the cache set from cache api

Hello, I’m Allie!•3/24/24, 11:08 PM

That can still be cleared via the API, if that is what you mean?

bun•3/24/24, 11:09 PM

yes true

bun•3/24/24, 11:09 PM

idk what you thought otherwise

SSilent i have to add type definitions or it manually or? did not mean to ping&*

Walshy•3/24/24, 11:11 PM

yeah you'll want to declare a module like:

declare module '*.html' {
  const value: string;
  export default value
}

declare module '*.html' {
  const value: string;
  export default value
}

deleted_user_97ffca1b58d4•3/25/24, 1:59 AM

I have a project in which I have to have a WebSocket connection on 24/7, watching for a (rare) event. If that event happens, I need to run some code, then restart the WebSocket and watch for the event again.

Is Workers suitable for this? From my understanding, there's no limit on wall time, so this would be feasible as long as the total CPU time stays under 10ms

bun•3/25/24, 3:53 AM

why is ctx.cf.city available on pages but not in workers ?

Walshy•3/25/24, 5:07 AM

It's available in both - https://reqinfo.walshy.dev/

If I'm worried about a Worker getting spammed (not a strong concern, just a thought), is there a way

Similar Threads

Similar Threads

Similar Threads