I'm using the

itty-router

for routing

typescript
export default {
  fetch: async (request: Request, env: Env, context: Context) => {
    context.slack = new Slack(env.SLACK_APP_BOT_TOKEN);
    context.waitUntil(router.handle(request, env, context));
  },
};

I have a following routes table: 1.

api.example.com/path-a/*

->

worker-a

2.

api.example.com/path-b/*

->

worker-b

3.

api.example.com/path-c*

->

worker-c

Every worker has a following setting in `worker.toml`:

[env.production]
name = "worker-a"
route = { pattern = "api.example.com/path-a/*", zone_name = "example.com" }

I fail to access these workers over HTTP. Looks like I'm missing something, help me out

My worker wrangler TOML file has following settings:

[[env.production.services]]
binding = "workerA"
service = "worker-a"

[[env.production.services]]
binding = "workerB"
service = "worker-b"

Both services are called like that:

const workerAResponse = await env.workerA.fetch(
  new Request(new URL(`worker-a/${id}`, env.WORKER_A_DOMAIN)),
  {
    headers: request.clone().headers,
  }
);


if (!workerAResponse.ok) { return workerAResponse }

const { public_id } = await workerAResponse.json();

const workerBResponse = await env.workerB.fetch(
  new Request(new URL(`worker-b/${public_id}`, env.WORKER_B_DOMAIN)),
  {
    headers: request.clone().headers,
  }
);

if (!workerBResponse.ok) { return workerBResponse; }

Request to

worker-a

is successful, however a request to` worker-b` is canceled, as

wrangler tail

says:

GET <some-url> - Canceled @ 4/14/2023, 5:03:23 PM

What can possibly go wrong about it? And why worker-b response is canceled whereas worker-a (both are set up literally the same) works just fine. By the way, worker responds with

200

, but due to the

outcome: "canceled"

it is not taken into account ;/

We are kicking the tyres of the Cloudflare ecosystem. We're using Terraform to stand up some infrastructure: - a KV store - a number of workers that have the KV store bound to them We can successfully deploy the workers but they come with a automatically created route that is not under the control of Terraform e.g.

delete_todo_dev.somedomain.workers.dev    somedomain.workers.dev

This route is disabled by default and can only be enabled in the UI. We are trying to enable the routes in Terraform so that we output the route's URL but as we didn't create the route, we can't interact with it. How is a Terraform project meant to create and manage routes programmatically? If we use the Wrangler tool, then the route gets enabled on creation but using Terraform it doesn't. We'd rather use Terraform than use Wrangler in a shell script, for obvious reasons. Any best practice advise is very welcome.

If I were to accept image uploads through a service, which then itself uploads the image to Cloudflare Images, would it be more suitable to do this through a second worker service, outside of my main API, or would it not matter in terms of networking?

i'm writing a worker / pages function and am getting a typescript error when i explicitly import a type into my .ts file. this code works fine and shows now errors:

export const onRequest: PagesFunction = async (context) => {
  return new Response( 'foobar' )
}

while this code causes a typescript error:

import type {
  PagesFunction,
} from '@cloudflare/workers-types/experimental'

export const onRequest: PagesFunction = async (context) => {
  return new Response( 'foobar' )
}

the error i'm getting is:

Type '(context: EventContext<unknown, any, Record<string, unknown>>) => Promise<Response>' is not assignable to type 'PagesFunction'.
  Type 'Promise<Response>' is not assignable to type 'Response | Promise<Response>'.
    Type 'Promise<Response>' is not assignable to type 'Promise<import(".../node_modules/@cloudflare/workers-types/experimental/index").Response>'.
      Type 'Response' is missing the following properties from type 'Response': webSocket, cf

i tried to ignore the error and delete the

import type

bits, but it's nagging at me and i can't let it go. 🙂 any thoughts or advice? has anyone else run into the same problem?

i wasn't sure if i should post this in pages-help or workers-help since it pertains to KV binding for local pages dev environments, so i'm just posting a link to the post i created over in pages-help: https://discord.com/channels/595317990191398933/1096910816999112714 this has been driving me mad the past 24+ hours, so any tips, pointers, or insights are appreciated! 🙂

I'm trying to create an integration with our CRM, however the application doesn't allow generating a set api key I can continue using, it requires users to log in using OAuth2. I'm using the service to process incoming orders sent to a webhook, and automatically enter them into the system. This means there's a highly irregular and unpredictable input. My question is, how do I keep refreshing the token, to make sure it doesn't expire (and require user interaction)?

app.get("/download", async (ctx) => { const address = ctx.req.query("address"); try { if (address) { const url = new URL(decodeURIComponent(address)); url.protocol = "https"; const { readable, writable } = new TransformStream(); url.protocol = "https"; const result = await fetch(url, { headers: { ...ctx.req.headers, redirect: "follow", "User-Agent": "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/112.0.0.0 Safari/537.36", }, }); let contentLength = result.headers.get("Content-Length"); let response = new Response(result.body); response.headers.append("Cache-Control", "no-cache"); response.headers.append("content-type", "audio/mpeg"); if (contentLength) { response.headers.append("Content-Length", contentLength); } response.body?.pipeTo(writable); return new Response(readable, response); } return ctx.newResponse("Not Found", 404); } catch (error) { console.log(error); return ctx.newResponse("Internal Server Error", 500); } }); I have this script that i can use to proxy a audio file from a url, how can i cache this audio file in cache? so i don't have to ping the server?

Hi. I'm using workers with the WebCrypto API for hashing, encryption and token auth. My main concern is though, is how slow PBKDF2 is. Should I be concerned about this? I'm currently hitting around ~20ms on average, with SHA-512 and 100k iterations. That does vary a bit though, and the main point of the worker is to hold encrypted content, of which could be 50mb in size or more for each individual thing.

bsically i'v brought a domain on cloudflare called cvrsocial.uk so i'v uploaded and deployed so i'm getting error could someone please elaborate on this also on that domain its giving me host error?

I’m seeing some very confusing cache-api behaviour on a worker of mine. I’ve set cache-control headers on a .css asset to be max-age=0, no-cache, but the worker is still caching the request. I can see this because the age response header is a non-zero value and the cf-cache-status says HIT. Is there any way I can turn off this automatic asset caching/any info of what’s going on under the hood?

Hello, I bought a domain name yesterday and a VPS on the same website : 1984.hosting, but I started to setup my domain with cloudflare, the nameservers has been updated, but when I go on my domain, it give me 1984.hosting, here is a screen of my DNS : ( sorry for the tags, idk what to put, i'm new on this server )

That's definitely doable, yes

Hello, I'm a learner of Cloudflare Worker 🙂 I faced some problems in a tutorial, so kindly help me. Thank you in advance. Detailed steps of my attempts are written below. MY GOAL: To connect my GoogleAnalytics and Clarity tags to my website. My site is hosted on root domain. (i.e. https://example.com) WHAT I UNDERSTOOD: Read some articles and found that it's better for my case to use Route matching (not Custom domain) because I just want to modify the html a little bit and want to apply it to an existing domain. NOW I'M TRYING: 1. Use a Worker like a proxy to modify HTML itself. (inject some script tags) → it didn't work. (details are in this post: https://community.cloudflare.com/t/cannot-trigger-worker-under-my-routing-googleanalytics-and-clarity-tags-with-htmlrewriter/498763?u=iam.nao.ito) 2. So, back to the basic. I tried to follow a tutorial: https://developers.cloudflare.com/workers/examples/alter-headers/ to review the basic rule. 3. Create a worker by following the tutorial 1. Create a new service worker on Cloudflare UI (dashboard) 2. Copy and paste the tutorial code in the

Quick edit

section of the worker. 3. Confirm that the header has changed on the preview panel. 4. Save and deploy 5. Create a new Route settings for the worker (pattern: example.com/, zone: example.com) 6. access to

https://example.com

→ The page is of course available like before, but the header is not changed like what was in the preview window (step 3-3). Also, the number of request didn't increase (it seems that my access is not routed to the worker). Is there any step I missed? Or, anyone faced / solved the same problem? Thank you.

can someone please pleaasae help i'm now worrying i want this worpress up and running my site is now showing i have DNS records correct as well can some please help

My understanding is that Wrangler will use esbuild to bundle any NPM packages upon deploy. I was trying to use just one package (probe-image-size) and can't get this to happen. I was developing locally and got errors then I tried the command to do a dry run locally and when I did so it still didn't bundle it... Is it possible to use NPM packages? What am I missing? Thank you!

Hi all. Is there any simple tutorial that shows how to build a chat app by using various CF services (e.g., Workers, KV, R2, etc) ? I found a chat app deployment tutorial but it's not that useful as it only contains front-end code. Thank you, Cloudflare. Your products are amazing! If only there were more docs available.

Hello, I am using two workers where one is working as a gateway to validate authentication and does a request to itself as a fallthrough (cf handlers first request, internal request goes to DNS origin) lets call this worker A Calling this worker via service binding from worker B results in a 200 response (using logging i can see that both the worker got the request with a 200 and the origin server logging results a 200) But this result is unreadable since its telling me the following: "This ReadableStream is disturbed (has already been read from), and cannot be used as a body." when i try to read the body. When doing a request from the browser to Worker A everything is working as expected. Path: Worker B -> Worker A -> Worker A (internal dns to kubernetes cluster in azure) I am wondering why the request is unreadable, is this a bug ? i expected the JSON result in the response but all i can see is the status code 200 and eveything else seems empty.

Hi, Studying what could fit in the well-known "10ms CPU Time", I was surprised by some "discrepancies": - the same code in a worker is going to be (statistically) slightly faster than in a page-function, - that same function - if the only one - is going to be (statistically) slightly faster than grouped with others even non used. Of course, it's never much, but enough to make it jump over the 10ms! Is that logical, even expected? Anyone has done some real extensive tests?

I'm trying to write a worker in Rust with the worker crate. I'm using the fetch API via the service worker global scope:

async fn request_sendgrid(api_key: &str, data: String) -> Result<u16, NetworkError> {
    let mut opts = RequestInit::new();
    opts.method("POST");
    opts.mode(RequestMode::Cors);
    opts.body(Some(&JsValue::from_str(&data)));

    let request = FetchRequest::new_with_str_and_init(&"https://api.sendgrid.com/v3/mail/send", &opts)?;

    request
        .headers()
        .set("Authorization", &format!("Bearer {}", api_key))?;
    request.headers().set("Content-Type", "application/json")?;

    let worker_global_scope = js_sys::global().dyn_into::<web_sys::ServiceWorkerGlobalScope>()?;
    let resp_value = JsFuture::from(worker_global_scope.fetch_with_request(&request)).await?;
    assert!(resp_value.is_instance_of::<FetchResponse>());
    let response: FetchResponse = resp_value.dyn_into()?;

    Ok(response.status())
}

The problem now is when running this, fetching the WASM fails with the above error at `build/worker/shim.mjs:181:20`:

function fetch(req, env, ctx) {
  const ret = wasm.fetch(addHeapObject(req), addHeapObject(env), addHeapObject(ctx));
  return takeObject(ret);
}

I'm trying to create a custom dashboard for my worker. Looking at the sum fields in the documentation I can't find what the units of the responses are. Can anyone assist? E.g. what is the unit for

wallTime

here?

working on getting a email worker to put a received email into a r2 bucket so far it works sorta. it creates the object but it empty. if i used

message.headers.get('from')

as the object it puts it in the r2 bucket as expected with the content being the sender, but if i use

message.headers.get('body')

it fails as if the object was empty. question is what are the email headers and am i right to think that 'body' is not one of them. i can't seem to find the doc with all the possible headers. if body isn't the header for the email content is there one?

Good day, using the example for allow list, the email trigger test will work properly. Yet sending an email to that route results in the reject message. I have validated the email address several times. Is there something trivial I am missing here? I did alter the reject message to make sure it was the worker sending the bounce.

Is there an API to create a worker? I wasn’t able to find it in the documentation. Specifically, I’m setting up a form to create redirects. * User visits my site and submits a form. * The values of the form’s input field are used in the creation of a new worker. * I would then attach this new worker to a route on my domain. I see the Worker Route API, but I don’t see a “create new worker” API. Am I missing something?

Https connection does not work properly, due unrecognized certificate.

anteac: I've read up all the docs but im not sure if those things provide a near kubernetes like cloud worker restarts or wehatever it needs to do to keep going, serverless., or microservices. idk what my usecase even is, serverless or microservices? Plan is running nextjs website and seperate API server on edgerouter mode on CF workers, coupled with some postgres supported DB and lastly a graphql server that a discord.js based bot consumes. theorectly possible all on cloudflare workers/pages, but I assumed you had to have some software to keep an eye on all the services to provide redundancy and selfhealing if needeed like kube. Like what am i missing here? Whats the best setup if I were to put everything in CF workers/pages? Do I even need a service discovery type thing, or do i need like a API gateway a lá Kong or something like these things https://github.com/envoyproxy/envoy or https://github.com/istio/istio or https://github.com/solo-io/gloo? anyone have experience with organizing something very similar recently? and if i need any of those api gateways?. I assume terraform would do the trick but I cant find any info on that combinened with edge workers only ,,,

I am struggling with ascertaining the type of

env

,

env.dispatcher

, and the result of

env.dispatcher.get(someString)

. Can anyone point me in the right direction? I was not able to find anything in

@cloudflare/worker-types/index.d.ts

.... Also for

env

itself, I kind of just type it like this:

type Env {
  dispatcher: any;
}

The relevant documentation: https://developers.cloudflare.com/cloudflare-for-platforms/workers-for-platforms/ It does not contain any TS 😦 Quick search on GH returns this: https://github.com/drivly/platform.workers.cf/blob/8fbdd76549b58414528611f75a04065fb1f49b12/src/env.ts#L9 Is that the recommended way to type this?

Did Cloudflare change request headers being sent from a worker while using fetch? Seems like my workers started to get CORS issues after the weekend all of a sudden.

Any explanation as to why workers logs don't work there and when this issue should be resolved? Thanks!