On free workers, that's probably the best strategy. If you have the paid plan though, a better optio

Z

zegevlierOP•9/7/24, 3:48 PM

If you are on the free plan and storing the tokens in KV, watch out that there aren't two things writing to it at the same time. Since KV is eventually consistent, another worker running might not see the most up-to-date value and might update the value incorrectly.

Zzegevlier On free workers, that's probably the best strategy. If you have the paid plan th...

J

jjjrmy•9/7/24, 10:47 PM

so is your suggestion to save the Oauth token in a durable object? I haven't really used them before.

AAngusMa If I call a service binding worker several times in a worker, each call can make...

A

AngusMa•9/8/24, 2:56 AM

Can I increase sub requests limit by calling them in another worker?

R

Rares•9/8/24, 9:30 AM

Hey, is

Object.groupBy

Object.groupBy

Object.groupBy

Object.groupBy avaliable inside the worker? it works locally but i want to be sure that it works when deployed

R

Rares•9/8/24, 10:16 AM

nice

R

Rob M.•9/8/24, 2:02 PM

Guys, I'm getting:

You are logged in with an API Token. Unset the CLOUDFLARE_API_TOKEN in the environment to log in via OAuth.

when trying to exec a command with the API token in .env file

R

Rob M.•9/8/24, 2:03 PM

It worked fine previously (wrangler 3.75.0)

R

Rob M.•9/8/24, 2:03 PM

Any thoughts?

M

majestic•9/9/24, 8:02 AM

Hi, can worker implement service to make request to another service that only receives request from static ip ?

I

Isak•9/9/24, 10:45 AM

Hi, I read somewhere that service bindings might become free because they're always loaded on the same machine. Any updates on this?

IIsak Hi, I read somewhere that service bindings might become free because they're alw...

E

Erisa•9/9/24, 12:13 PM

they became free recently: https://developers.cloudflare.com/workers/platform/pricing/#service-bindings

Requests made from your Worker to another worker via a Service Binding do not incur additional request fees. This allows you to split apart functionality into multiple Workers, without incurring additional costs.

M

Murder Chicken•9/9/24, 12:59 PM

Is it possible to send key metadata that will be part of the events passed to a tail worker? For instance an application request UUID? Obviously, we can pick up the

console.log()

console.log()

console.log()

console.log() and

exceptions

exceptions

exceptions

exceptions but is there a way to add a metadata field that's added to the

event

event

node of what's passed to the

tail()

tail()

tail()

tail() method in a Tail worker?

Original message was deleted

C

Chaika•9/9/24, 1:23 PM

There's a lot of magic. Did you try on your workers.dev or on a custom domain?

Original message was deleted

C

Chaika•9/9/24, 1:25 PM

I would try your workers.dev, Custom Domains can have things that effect outbound fetches like cache rules, I believe Transform Rules would also touch. Your SSL/TLS Mode also can transparently change the encryption regardless of what you specify in the fetch (ex: Flexible or Off will transparently downgrade https:// to http)

P

Ping for toast•9/9/24, 1:31 PM

is it possible to improve cold start times

P

Ping for toast•9/9/24, 1:31 PM

cold starts see ~500ms, subsequent requests are ~150ms-250ms

E

Erisa•9/9/24, 1:31 PM

make the worker smaller?

EErisa make the worker smaller?

P

Ping for toast•9/9/24, 1:32 PM

How much does size impact it? It is a large application and there are some large WASM dependencies (mostly for OG image generation)

P

Ping for toast•9/9/24, 1:33 PM

GZ size is 2mb I think

E

Erisa•9/9/24, 1:33 PM

larger workers have longer cold starts, its one of the main causes but isnt the only cause (theres also v8 optimisations for cpu-heavy workers as leo says)

EErisa larger workers have longer cold starts, its one of the main causes but isnt the ...

P

Ping for toast•9/9/24, 1:34 PM

is there something I can do to fix that? I assume my main option is to seperate the worker into smaller pieces?

P

Ping for toast•9/9/24, 1:35 PM

og is 27% of the bundle

E

Erisa•9/9/24, 1:35 PM

splitting would work if each execution only needed some or none of the split workers, wouldnt be very useful if you needed to call all of them anyway

E

Erisa•9/9/24, 1:35 PM

just generally the only advice in this case is to make the worker smaller, how to do that will depend on the exact use-case

EErisa splitting would work if each execution only needed some or none of the split wor...

P

Ping for toast•9/9/24, 1:36 PM

This would be the case. OG image generation is just needed when another website requests it

E

Erisa•9/9/24, 1:37 PM

then yeah try splitting OG image gen into its own worker

P

Ping for toast•9/9/24, 1:37 PM

I'll try that ig

E

Erisa•9/9/24, 1:37 PM

test whether that helps and iterate accordingly

E

Erisa•9/9/24, 1:37 PM

you should connect to the split worker over a service binding to remove any network latency from the worker-to-worker fetch

EErisa you should connect to the split worker over a service binding to remove any netw...

P

Ping for toast•9/9/24, 1:38 PM

Could I also do workers routes? If I had one worker pointing to

/api/*

/api/*

/api/*

/api/* and another one pointing to

/api/og/*

/api/og/*

/api/og/*

/api/og/*

PPing for toast og is 27% of the bundle

P

Ping for toast•9/9/24, 1:39 PM

One bundle size fantasy (unrelated to OG generation) I've had for years is macros in bundlers

E

Erisa•9/9/24, 1:39 PM

oh rite you said the website is requesting the worker, in that case yeah you can do routes like that

E

Erisa•9/9/24, 1:39 PM

its if you wanted your main worker to act on the response that youd have to use service binding since same-zone fetch doesnt work

P

Ping for toast•9/9/24, 1:40 PM

ye

E

Erisa•9/9/24, 1:40 PM

but if its just a different api endpoint by all means use routes like that

P

Paolo•9/9/24, 4:25 PM

does a worker created with the dashaboard support ES modules? Or can that be done only via wrangler? I see the wrangler.toml file at creation but I can't edit it. Can it be done after creation?

M

Mike D.•9/9/24, 6:24 PM

https://github.com/cloudflare/workers-sdk/issues/6657 having issue with typescript

GitHub

🐛 BUG: Uncaught SyntaxError: Unexpected token '<' for embedded type...

Which Cloudflare product(s) does this pertain to? Wrangler What version(s) of the tool(s) are you using? 3.75.0 wrangler What version of Node are you using? v22 pre What operating system and versio...

M

Mike D.•9/9/24, 6:26 PM

https://github.com/meta-introspector/mina-snarky-o1js-zkapp-introspector/actions/runs/10778636802/job/29890420745#step:4:61 this one in github actions fails because it cannot find the source

GitHub

moving to ui dir · meta-introspector/mina-snarky-o1js-zkapp-introsp...

zkIntrospector: A Self-Aware zkApp Framework for Gödel Number Reflection - moving to ui dir · meta-introspector/mina-snarky-o1js-zkapp-introspector@a6f3f9a

M

Mike D.•9/9/24, 7:05 PM

i will rename it if it helps

M

Mike D.•9/9/24, 7:25 PM

that did not help, also o1js needs threads so i might have to go to another solutions

G

garb•9/10/24, 2:25 AM

Hey guys, is anyone else's audit throwing a high severity error with cloudflare/vitest-pool-workers & wrangler.

G

garb•9/10/24, 2:26 AM

# npm audit report

path-to-regexp  0.2.0 - 7.2.0
Severity: high
path-to-regexp outputs backtracking regular expressions - https://github.com/advisories/GHSA-9wv6-86v2-598j
No fix available
node_modules/path-to-regexp
  wrangler  <=0.0.0-kickoff-demo || >=0.0.7
  Depends on vulnerable versions of path-to-regexp
  node_modules/wrangler
    @cloudflare/vitest-pool-workers  *
    Depends on vulnerable versions of wrangler
    node_modules/@cloudflare/vitest-pool-workers

3 high severity vulnerabilities

# npm audit report

path-to-regexp  0.2.0 - 7.2.0
Severity: high
path-to-regexp outputs backtracking regular expressions - https://github.com/advisories/GHSA-9wv6-86v2-598j
No fix available
node_modules/path-to-regexp
  wrangler  <=0.0.0-kickoff-demo || >=0.0.7
  Depends on vulnerable versions of path-to-regexp
  node_modules/wrangler
    @cloudflare/vitest-pool-workers  *
    Depends on vulnerable versions of wrangler
    node_modules/@cloudflare/vitest-pool-workers

3 high severity vulnerabilities

# npm audit report

path-to-regexp  0.2.0 - 7.2.0
Severity: high
path-to-regexp outputs backtracking regular expressions - https://github.com/advisories/GHSA-9wv6-86v2-598j
No fix available
node_modules/path-to-regexp
  wrangler  <=0.0.0-kickoff-demo || >=0.0.7
  Depends on vulnerable versions of path-to-regexp
  node_modules/wrangler
    @cloudflare/vitest-pool-workers  *
    Depends on vulnerable versions of wrangler
    node_modules/@cloudflare/vitest-pool-workers

3 high severity vulnerabilities

# npm audit report

path-to-regexp  0.2.0 - 7.2.0
Severity: high
path-to-regexp outputs backtracking regular expressions - https://github.com/advisories/GHSA-9wv6-86v2-598j
No fix available
node_modules/path-to-regexp
  wrangler  <=0.0.0-kickoff-demo || >=0.0.7
  Depends on vulnerable versions of path-to-regexp
  node_modules/wrangler
    @cloudflare/vitest-pool-workers  *
    Depends on vulnerable versions of wrangler
    node_modules/@cloudflare/vitest-pool-workers

3 high severity vulnerabilities

G

garb•9/10/24, 2:27 AM

Wrangler is on version 3.75.0.

S

sathoro•9/10/24, 4:36 AM

anyone know if it is possible to use fetch with a proxy now? I tried

https-proxy-agent

https-proxy-agent

https-proxy-agent

https-proxy-agent but getting

TypeError: Cannot convert object to primitive value

TypeError: Cannot convert object to primitive value

TypeError: Cannot convert object to primitive value

TypeError: Cannot convert object to primitive value when calling

new HttpsProxyAgent(proxyUrl)

new HttpsProxyAgent(proxyUrl)

new HttpsProxyAgent(proxyUrl)

new HttpsProxyAgent(proxyUrl)

S

sathoro•9/10/24, 4:41 AM

another attempt with a different package, no error but this just ignores the proxy

import { createFetch } from "node-fetch-native/proxy";

const fetch = createFetch({ url: proxyUrl });
await fetch(targetUrl);

import { createFetch } from "node-fetch-native/proxy";

const fetch = createFetch({ url: proxyUrl });
await fetch(targetUrl);

import { createFetch } from "node-fetch-native/proxy";

const fetch = createFetch({ url: proxyUrl });
await fetch(targetUrl);

import { createFetch } from "node-fetch-native/proxy";

const fetch = createFetch({ url: proxyUrl });
await fetch(targetUrl);

S

sathoro•9/10/24, 4:43 AM

I'm guessing whatever it does behind-the-scenes just gets pollyfilled back to CF's regular fetch function so it won't be possible

J

Jess•9/10/24, 4:53 AM

Hello my name is jess and I would like to know if there is any advise for captcha error when trying to log in using citrix?

G

garb•9/10/24, 5:13 AM

I submitted a forum in #workers-help

On free workers, that's probably the best strategy. If you have the paid plan though, a better optio

Similar Threads

Similar Threads

Similar Threads