Your Wallet, basically

HHello, I’m Allie!Your Wallet, basically

Hacker Factor•10/24/24, 3:06 PM

That's not helping. With $5, I get 30 million requests per month. That's good. What if I need 60 million? With the $5 plan, I pay something like $0.30 per million, so that's an additional $9 for that month. What is the request limit for the Pro plan before getting overages? (If one of my services is getting 60 million requests per month, then wait until I bring over the other three services...)

HHacker Factor That's not helping. With $5, I get 30 million requests per month. That's good. W...

Hello, I’m Allie!OP•10/24/24, 3:09 PM

Basically, with the paid plan, you get the 10 million request/30 million CPU milliseconds a month included, then pay $0.3 per additional million requests, and $0.02 per additional million CPU milliseconds. That's it.

Hello, I’m Allie!OP•10/24/24, 3:10 PM

You can scale to billions of requests with the exact same pricing

Hello, I’m Allie!OP•10/24/24, 3:10 PM

There isn't a cap for "number" of requests at that point

Hacker Factor•10/24/24, 3:19 PM

Got it. Thank you.

dave•10/24/24, 6:30 PM

yarn create v1.22.22
warning package.json: No license field
[1/4] 🔍  Resolving packages...
[2/4] 🚚  Fetching packages...
error @achrinza/node-ipc@9.2.6: The engine "node" is incompatible with this module. Expected version "8 || 9 || 10 || 11 || 12 || 13 || 14 || 15 || 16 || 17 || 18 || 19". Got "22.9.0"
error Found incompatible module.

yarn create v1.22.22
warning package.json: No license field
[1/4] 🔍  Resolving packages...
[2/4] 🚚  Fetching packages...
error @achrinza/node-ipc@9.2.6: The engine "node" is incompatible with this module. Expected version "8 || 9 || 10 || 11 || 12 || 13 || 14 || 15 || 16 || 17 || 18 || 19". Got "22.9.0"
error Found incompatible module.

What causes an old version of @achrinza/node-ipc@achrinza/node-ipc to be used when when I run yarn create cloudflare@latest workflows-starter-hello-world-1 --template "cloudflare/workflows-starter"yarn create cloudflare@latest workflows-starter-hello-world-1 --template "cloudflare/workflows-starter"? There is a newer version of @achrinza/node-ipc@achrinza/node-ipc that supports version 22 of node, so I don't get why it's insisting on trying an older version.

Avi•10/24/24, 7:18 PM

is anyone aware of ratelimiting or other limits that would cause

fetch()

fetch()

requests initiated by a DurableObject to sometimes take exactly 10 seconds longer than they otherwise would?

Avi•10/24/24, 7:19 PM

typically these requests take < 300ms. sometimes however, they take 10000+(<300)ms

Avi•10/24/24, 7:19 PM

i.e. we never see 2 seconds, 3 seconds, 9 seconds... just < 300ms and ~10.1 - 10.3 seconds

Avi•10/24/24, 7:20 PM

this happens for requests to multiple completely unrelated services

Avi•10/24/24, 7:20 PM

so i'm wondering if there's some sort of limit happening we're not aware of

Advany•10/24/24, 8:21 PM

deploying is giving me: ✘ [ERROR] response.errors is not iterable

Advany•10/24/24, 8:22 PM

AAdvany deploying is giving me: ✘ [ERROR] response.errors is not iterable

Chaika•10/24/24, 8:24 PM

try again?

Chaika•10/24/24, 8:24 PM

I just got one of those from a different api:

Non-Success with no errors, response body: { "code": 1003, "error": "auth.jwt_headers_invalid" }

but it succeeded on retry, hmmm

AAdvany Click to see attachment

Frank•10/24/24, 8:24 PM

same here

Frank•10/24/24, 8:32 PM

worked after 3 retries

CChaika I just got one of those from a different api: > Non-Success with no errors, resp...

Advany•10/24/24, 8:34 PM

jup, working again!

Brett Willis•10/25/24, 5:50 AM

Hi team. How exactly is the open source

workerd

workerd

related to the production runtime?
Is

workerd

workerd

running in prod with extras wrapped around it?
Or does prod apply closed source patches to

workerd

workerd

?
Or is

workerd

workerd

just a similar but separate codebase?
I ask because if I look at a particular implementation detail in

workerd

workerd

e.g. polyfills or how the event loop works is that how it is in prod?
Also if I were to contribute to

workerd

workerd

(if allowed) would that make it to prod?

Walshy•10/25/24, 5:54 AM

Yes, the Workers runtime is just an extension to workerd with the internal Cloudflare bits. Most functionality is within workerd, just things like limiting, internal routing etc are in the closed source runtime

and yes, if you contribute to workerd, it'll make it to prod. We update the workerd module frequently

Constantin•10/25/24, 7:41 AM

howdy, is it possible to use my own sass platform custom domain for dcv TXT record?

CConstantin howdy, is it possible to use my own sass platform custom domain for dcv TXT reco...

fry69•10/25/24, 7:58 AM

You _acme_acme CNAME (from your sass domain/platform) needs to point to Cloudflare for them to be able to issue certificates.

fry69•10/25/24, 8:00 AM

This effectively disables any ACME challenge on your own behalf for this host, but apparently this is what you want?

fry69•10/25/24, 8:02 AM

syph•10/25/24, 2:10 PM

Hello , does anyone have a repo or a tutorial they can share on how to implement social login securely with cloudflare workers, or is it possible at all or do i have to have a server. thanks

Ssyph Hello , does anyone have a repo or a tutorial they can share on how to implement...

fry69•10/25/24, 2:15 PM

Seems possible with e.g. GitHub -> https://github.com/gr2m/cloudflare-worker-github-oauth-login

ac•10/25/24, 3:02 PM

I'm having an issue again with worker builds just getting stuck at "Queued" forever (first discussed workers-and-pages-discussions). Only pushed changes for the single worker, it immediately changed to "Queued" (no other workers in the account were triggered, which is correct). Just stuck on "Queued" forever....

ac•10/25/24, 3:05 PM

Nevermind I guess, it fixed itself after about 10 minutes..

Not really clear what it was waiting on since no builds were running in my account. Is there a global build queue?

Aac Nevermind I guess, it fixed itself after about 10 minutes.. 🤔 Not really clear ...

fry69•10/25/24, 3:08 PM

I assume there is a global distribution step inside the build process which can things slow down if the acknowledgements do not come back in time, but 10 minutes sounds excessive. Then again I have no real clue.

fry69•10/25/24, 3:10 PM

Also if those build processes happen in the CI or via pull from Cloudflare from GitHub 10 min (as an outlier) still seems fine. I know mandatory check pipelines which are measured in hours.

Raiyan•10/25/24, 3:16 PM

is there any way to increase worker size limit without upgrading to paid plan? I've a very small application, might get at most 100 req/day, most probably used by only me, but I couldn't compact it to 1MB limit, and getting paid plan just for the size limit doesn't seem feasible

RRaiyan is there any way to increase worker size limit without upgrading to paid plan? I...

fry69•10/25/24, 3:18 PM

Use multiple worker in a chain?

Ssyph Hello , does anyone have a repo or a tutorial they can share on how to implement...

Raiyan•10/25/24, 3:19 PM

you can use lucia, and if you need a bit more control, you can use arctic, created by the same person as lucia, arctic is a package that makes it easy to generate oauth state and callback links with many providers

fry69•10/25/24, 3:20 PM

Also if you need so few requests and only by you, how about a simple dedicated VPS box?

Ffry69 Use multiple worker in a chain?

Raiyan•10/25/24, 3:20 PM

that would be unmaintainable, of course I can split the functionality to different workers and use rpc, but that would be overcomplication, and my worker size isn't even very big, it is 1.2MB, at the end of the project, it may become 2MB at most, but not more than that

Ffry69 Also if you need so few requests and only by you, how about a simple dedicated V...

Raiyan•10/25/24, 3:21 PM

$4 bucks against free, getting $5 paid plan would be better for having access to much more things

RRaiyan is there any way to increase worker size limit without upgrading to paid plan? I...

Chaika•10/25/24, 3:22 PM

well the root answer to your question is no, although I would make sure you're looking at the gzipped size and not the raw size, might be possible to strip something from your worker as well, 1.2mb is kind of large for a worker

fry69•10/25/24, 3:22 PM

Those $4 give you a small datacenter these days (compared to a few decades back), that is a lot of concentrated (albeit also centralized) compute and space.

CChaika well the root answer to your question is no, although I would make sure you're l...

Raiyan•10/25/24, 3:25 PM

it is 248KB if I calculate gzip

RRaiyan it is 248KB if I calculate gzip

Chaika•10/25/24, 3:26 PM

the gzipped size is what matters. If you do wrangler deploy it tells you the size. Did it still fail to deploy or something?

Raiyan•10/25/24, 3:26 PM

I haven't deployed it, I added a big ass library and it added around 500KB to the bundle

Chaika•10/25/24, 3:27 PM

you can just dump it to a folder like wrangler deploy --outdir bundled/ --dry-runwrangler deploy --outdir bundled/ --dry-run, and should tell you the raw and gzipped

Raiyan•10/25/24, 3:31 PM

I used this command to get the exact size which was 1.2MB, then used esbuild as wrangler uses esbuild to bundle to get the meta file and from there, I saw the library adds around 500KB, but using gzip utility, I got 248KB to be the file size

Raiyan•10/25/24, 3:32 PM

there is now static hosting in workers, how that works?

Raiyan•10/25/24, 3:32 PM

is it bundled with the worker itself?

RRaiyan I used this command to get the exact size which was 1.2MB, then used esbuild as ...

Chaika•10/25/24, 3:32 PM

Not sure I would trust the gzip utility, could be different compression levels and such, I would go with the dry run or actual deployment

RRaiyan is it bundled with the worker itself?

Chaika•10/25/24, 3:32 PM

completely separate

CChaika Not sure I would trust the gzip utility, could be different compression levels a...

Raiyan•10/25/24, 3:33 PM