Hello everyone, if I were to deploy some frontend project to Cloudflare, I would need to choose betw

Hello everyone, if I were to deploy some frontend project to Cloudflare, I would need to choose between Pages and Workers Static Assets. The recommendation are a little bit ambiguous to me, from what I can understand, we can use both and they'll eventually be the same.

But in the meantime, I should prefer Worker Static Assets over Pages, right ?

HHebilicious Hello everyone, if I were to deploy some frontend project to Cloudflare, I would...

Ashley•12/2/24, 9:35 PM

Probably, but there are some Pages-specific features that are not yet availlable for Static Assets (see https://developers.cloudflare.com/workers/static-assets/compatibility-matrix/). Generally though, they should converge over time is my understanding.

Yunfi•12/2/24, 11:34 PM

I know that wrangler use esbuild to bundle code. Is there a way i can pass additional flags/configurations to esbuild when dev or deploy?

IIsaac McFadyen What I'm wondering is if the actual billing data increased by 10x on your invoic...

MetaSky•12/3/24, 2:17 AM

This is a comparison of the October and September bills, from the bills, it is true that the cpu time calculation has been increased by 10 times, according to the current calculation method to calculate the cpu time in September, it is not possible to correspond to the September bill, I would like to confirm what kind of calculation is correct, is it the calculation method of the September bill, or the current calculation method of multiplying it by 10 times.

MMetaSky This is a comparison of the October and September bills, from the bills, it is t...

1984 Ford Laser•12/3/24, 4:32 AM

I'm not sure because I don't use all of CF's products, but could some other service be contributing to Worker CPU MS that isn't covered by WorkersOverviewRequestsAdaptiveGroups? Like did you start using something else that uses that same billing metric? Just speculation

11984 Ford Laser I'm not sure because I don't use all of CF's products, but could some other serv...

MetaSky•12/3/24, 4:46 AM

No, I've only deployed the workers script, and I'd like an official answer about the cpu time calculation.

MMetaSky No, I've only deployed the workers script, and I'd like an official answer about...

1984 Ford Laser•12/3/24, 4:46 AM

Have you filed a billing ticket? Discord probably won't get a satisfactory official response

11984 Ford Laser Have you filed a billing ticket? Discord *probably* won't get a satisfactory off...

MetaSky•12/3/24, 4:50 AM

Thanks a lot, I'll try to submit a billing ticket to https://community.cloudflare.com

MMetaSky This is a comparison of the October and September bills, from the bills, it is t...

Omar Abid•12/3/24, 5:43 AM

I am not sure I understand your question. It seems your usage went x10 and so did your billing costs?

OOmar Abid I am not sure I understand your question. It seems your usage went x10 and so di...

MetaSky•12/3/24, 7:26 AM

What I mean is that the cpu time data I queried with the October api matched the October bill,
but the cpu time data I queried with the same api for September didn't match the September bill,
if cloudflare x10s the cpu time calculation strategy in October and that's what happens.
Because I compared my September and October without changing the worker's script code.

GuillaumeLakano•12/3/24, 8:16 AM

Hi! We trying to move Pages to a « Static Assets » (Worker with Static Assets but without any Worker), because this seems that Pages will disappear at some point in the future and be replaced with this new product.

But both product doesn't have the same behavior for routing:

Imagine this structure:
/file.html
/folder/index.html

Pages:
« domain.tld/file.html » => status 200 and return /file.html
« domain.tld/folder/ » => status 200 and return /folder/index.html

Static Assets with html_handling = 'auto-trailing-slash' :

« domain.tld/file.html » => 307 to « /file » and the 2nd request return /file.html with status 200
« domain.tld/folder/ » => 200 and return /folder/index.html

Static Assets with html_handling = 'force-trailing-slash' :

« domain.tld/file.html » => 307 to « /file/ » and the 2nd request return /file.html with status 200
« domain.tld/folder/ » => 200 and return /folder/index.html

Static Assets with html_handling = 'drop-trailing-slash' :

« domain.tld/file.html » => 307 to « /file » and the 2nd request return /file.html with status 200
« domain.tld/folder/ » => 307 to « /folder » and the 2nd request return /folder/index.html with status 200

Static Assets with html_handling = 'none' :

« domain.tld/file.html » => 200 and return /file.html
« domain.tld/folder/ » => 404

How can we have the exact behavior of Pages, without any redirection 307 ?
May be we need a new option like html_handling = 'standard' ( whatever the name ) for acting like Pages and standard webservers ?

MMetaSky What I mean is that the cpu time data I queried with the October api matched the...

Omar Abid•12/3/24, 9:33 AM

Not changing the code doesn't mean that usage didn't change. If you are using DO/WebSockets, you might have some left resources that are burning cpu times. You'll need to watch your cpu consumption and find out where it's coming from.

OOmar Abid Not changing the code doesn't mean that usage didn't change. If you are using DO...

MetaSky•12/3/24, 9:57 AM

Thank you very much for your help, but that doesn't explain the difference in cpu time calculation between October and September!

kamama•12/3/24, 1:24 PM

Hi, may i know if this will keep produce cost?

import { connect } from 'cloudflare:sockets';

// if keep socket connection here will charge CPU time?
const socket = connect(gopherAddr);


export default {
  async fetch(req): Promise<Response> {
  }
}

import { connect } from 'cloudflare:sockets';

// if keep socket connection here will charge CPU time?
const socket = connect(gopherAddr);


export default {
  async fetch(req): Promise<Response> {
  }
}

Hugo•12/3/24, 2:13 PM

Dumb question:

I use the MongoDB Realm SDK to query my db from Workers, as the NodeJS SDK is not supported in Workers runtime.
Realm SDK is now deprecated.
Now what? There is just no way to use MongoDB in a Worker runtime anymore?

SuperHelpflare•12/3/24, 2:27 PM

Please do not post your question in multiple channels/post it multiple times per the rules at #

welcome-and-rules. It creates confusion for people trying to help you and doesn't get your issue or question solved any faster.

Leo•12/3/24, 2:44 PM

I’m really curious about how Workers with Assets try to match static assets first and then route to a worker if they don’t find a match. Is there an internal worker that handles the routing and matching upfront?

LLeo I’m really curious about how Workers with Assets try to match static assets firs...

GuillaumeLakano•12/3/24, 3:10 PM

From what I've understood, you can put a worker in front ( all requests uses the worker ) or without worker. All depend of how you setup your wrangler.toml ( « main = './src/index.js' » )
Then, you define in wrangler the folder of your static assets ( « assets = { directory = "./public/" } » )
But, unfortunately, I don't think there is a way to first try to return a static asset and then run the worker. There is a not_found_handler but only to set a SPA ( fallback to the /index.html ) or to set a 404 page

GGuillaumeLakano From what I've understood, you can put a worker in front ( all requests uses the...

texan•12/3/24, 3:18 PM

Very soon (in the next few hours) you'll be able to do this. You'll be able to indiscriminately run a user worker (even on a path that would have otherwise matched an asset) and dynamically fetch from the assets binding. Then you could do whatever you wanted to the assets (think, HTMLRewriter) in user worker code

texan•12/3/24, 3:19 PM

Also thanks, Lakano for the issue report, it's being floated around internally

ac•12/3/24, 7:03 PM

Should preview environments be working with Workers running Static Assets (like a Remix app)? Mine deploys properly, but it's not showing any of the changes

ac•12/3/24, 7:20 PM

In fact, it looks like it's showing super old versions of code. Nothing even remotely close to the actual "live" code (which is deployed via Workers Builds)

Jason Barry•12/3/24, 7:26 PM

does anyone have a working example of using

socket.startTls()

socket.startTls()

to upgrade a socket to be secure? I've followed these docs which for me works with

wrangler dev

wrangler dev

but not

wrangler dev --remote

wrangler dev --remote

. from a worker, I want to connect to an HTTP proxy with the HTTP CONNECT method, then perform TLS handshake, and then GET a resource from a target HTTPS site via the proxy connection

Constantin•12/3/24, 7:58 PM

how CPU Time Limit (ms) limit can be set via the wrangler.toml?

Walshy•12/3/24, 8:01 PM

https://developers.cloudflare.com/workers/wrangler/configuration/#limits

makisuo•12/3/24, 8:28 PM

I have a weird issue that in two of my projects dependencies dont seem to get installed anymore on deploy. Is that a known issue right now?

alotabits•12/3/24, 11:14 PM

Is there an exhaustive list of compatibility_date changes kept in one searchable place? The workers docs have a recent subset but I need to upgrade from an older date.

Constantin•12/4/24, 1:13 AM

you can see those inside dash

Constantin•12/4/24, 1:14 AM

worker / settings -> runtime (compatibility date)

CConstantin you can see those inside dash

Isaac McFadyen•12/4/24, 1:14 AM

The dash actually pulls from the same source as the docs so if any aren't in the docs they also won't be in the dash unfortunately.

Constantin•12/4/24, 1:15 AM

and here: https://developers.cloudflare.com/workers/configuration/compatibility-flags/#flags-history

Cloudflare Docs

Compatibility flags · Cloudflare Workers docs

Opt into a specific features of the Workers runtime for your Workers project.

Constantin•12/4/24, 11:17 AM

hei, can turnstile be used for custom hostnames?

Constantin•12/4/24, 11:18 AM

i want to add protection to my tenants contact forms, but can't see how to include these domains

dave•12/4/24, 1:33 PM

How do I extract the text inside

<RESULT>1.214</RESULT>

<RESULT>1.214</RESULT>

using

HTMLRewriter

HTMLRewriter

<?xml version="1.0" standalone="no" ?><RESPONSE><RESULT>1.214</RESULT></RESPONSE>

<?xml version="1.0" standalone="no" ?><RESPONSE><RESULT>1.214</RESULT></RESPONSE>

Ddave How do I extract the text inside `<RESULT>1.214</RESULT>` using `HTMLRewriter`? ...

dave•12/4/24, 2:38 PM

ah, got it working by using

.on("result")

.on("result")

Tom Austin•12/4/24, 2:54 PM

hi all, quick question. I've got a worker that interacts with the cache api however I could still see data being retrieved from the origin when it should be coming from the cache. I've now found that workers without a custom domain have no access to functional cache operations. I've now added a custom domain to my worker. Is my assumption that the issue was the missing custom domain and now that I've added it everything should work correct? Thanks

TTom Austin hi all, quick question. I've got a worker that interacts with the cache api howe...

lucidplot•12/4/24, 3:03 PM

you can see if a request was cached by looking at the response headers

Tom Austin•12/4/24, 3:04 PM

I'm seeing CF-Cache-Status HIT so I guess that's all good then

Shados•12/4/24, 3:18 PM

I can see from the release notes which version of v8 is used by Workerd/Cloudflare Workers. Is there a good resource to tell at a glance what JavaScript features are supported in a given v8 version?

Divanshu Chauhan•12/4/24, 4:07 PM

Hey guys, I don’t if this is the correct place to ask this question but I guess it relates closely to workers hence asking.

I wish to ask for a scenario (hypothetical):

Lets say I have my content in backblaze b2 (around 14tb, mostly photos, videos and zip files). I have around 100-120tb egress per month from b2 to users. How would bandwidth alliance work with this? If I do b2 -> cf worker -> user. What chargers can I expect from cloudflare? I know b2 would charge me for storage and egress between b2 and cloudflare is free because if bandwidth alliance but what about cloudflare -> user?

Original message was deleted

Constantin•12/4/24, 4:25 PM

any ideea how much does that cost?

Negative•12/4/24, 7:10 PM

Is it a bad idea to have a DO that subscribes to a Graphql stream indefinitely? It would be responsible for creating new DOs and broadcasting to them from results in the stream. Running 24/7

Negative•12/4/24, 7:11 PM

Basically proxying this one stream into many dedicated streams.

dave•12/4/24, 7:19 PM

Any ideas on how can use a Worker to pick the "closest" server city? e.g.

Miami

Miami

is the query, and

["NYC", "Toronto"]

["NYC", "Toronto"]

is the list of server cities.

dave•12/4/24, 7:20 PM

oh I can use lat/long

NNegative Is it a bad idea to have a DO that subscribes to a Graphql stream indefinitely? ...

Hard@Work•12/4/24, 7:24 PM

It's not a bad idea per-se? You just need some logic to handle reconnections in the event you are disconnected from the stream

rivage•12/4/24, 8:20 PM

It would be awesome if tailing Worker logs streamed logs as they were created, not when requests conclude

rivage•12/4/24, 8:26 PM

Or started steaming logs after walling >15s

rivage•12/4/24, 8:28 PM

Transparency into when queues call consumers would be a welcome addition as well

kcbrewron•12/4/24, 9:19 PM

Shout-out to @boristane presenting at AWS reinvent

codebam•12/4/24, 10:45 PM

can someone help me with @cf/black-forest-labs/flux-1-schnell

import { connect } from 'cloudflare:sockets'; // if keep socket connection here will charge CPU time? const socket = connect(gopherAddr); export default { async fetch(req): Promise<Response> { } }

Hello everyone, if I were to deploy some frontend project to Cloudflare, I would need to choose betw

Similar Threads

Similar Threads

Similar Threads