The client, as in an external customer not using CF? - Cloudflare Developers

© 2026 Hedgehog Software, LLC

More

Communities Docs About Terms Privacy

Chaika

CChaika The client, as in an external customer not using CF?

elpupper

E

elpupper•10/31/24, 8:38 PM

Yes

elpupper

E

elpupper•10/31/24, 8:39 PM

So the wildcard acts as a proxy

Chaika

C

ChaikaOP•10/31/24, 8:39 PM

Then you have to use CF For SaaS, which doesn't support wildcards full stop for non-enterprise

Chaika

C

ChaikaOP•10/31/24, 8:39 PM

Can only use Worker Routes and such when the domain/zone is in your account

Chaika

C

ChaikaOP•10/31/24, 8:40 PM

You can't just CNAME to a worker either if that's what you're thinking

Mozzy

M

Mozzy•10/31/24, 9:29 PM

Is there a good/native way of avoiding this behaviour in order to get a cache miss?

Mozzy

M

Mozzy•10/31/24, 9:31 PM

Currently, I'm just doing a simple cache with Hono middleware

export const cache = (name: string, minutes: number) =>
  honoCache({
    cacheName: name,
    cacheControl: `public, max-age=${minutes * 60}`,
  });

export const cache = (name: string, minutes: number) =>
  honoCache({
    cacheName: name,
    cacheControl: `public, max-age=${minutes * 60}`,
  });

Mozzy

MMozzy Is there a good/native way of avoiding this behaviour in order to get a cache mi...

Isaac McFadyen

I

Isaac McFadyen•10/31/24, 9:35 PM

In general this is a good way to cache-bust. It seems "hacky" but it's not really and is used quite frequently in a lot of sites actually.

Spunkie

S

Spunkie•10/31/24, 9:35 PM

So routes is not available on pages, only workers? Is that right or am I just blind?

Spunkie

SSpunkie So routes is not available on pages, only workers? Is that right or am I just bl...

Isaac McFadyen

I

Isaac McFadyen•10/31/24, 9:35 PM

That's correct. Pages is full-domain.

Isaac McFadyen

IIsaac McFadyen That's correct. Pages is full-domain.

Spunkie

S

Spunkie•10/31/24, 9:36 PM

:weicry:

Isaac McFadyen

IIsaac McFadyen In general this is a good way to cache-bust. It seems "hacky" but it's not reall...

Mozzy

M

Mozzy•10/31/24, 9:37 PM

I know it's helpful, but I would like the option to have stricter cache rules

Mozzy

M

Mozzy•10/31/24, 9:37 PM

Maybe a rewrite middleware before the caching?

Spunkie

SSpunkie :weicry:

Isaac McFadyen

I

Isaac McFadyen•10/31/24, 9:37 PM

You could write a Pages Function to proxy some routes to another backend but that would increase latency and always bill you for Function invocations.

Mozzy

MMozzy Maybe a rewrite middleware before the caching?

Isaac McFadyen

I

Isaac McFadyen•10/31/24, 9:37 PM

Oh I might have misunderstood. You want to disable that so it still hits the cache?

Mozzy

M

Mozzy•10/31/24, 9:37 PM

Exactly

Isaac McFadyen

I

Isaac McFadyen•10/31/24, 9:37 PM

If so there's a setting you can change in the Cloudflare dashboard.

Isaac McFadyen

I

Isaac McFadyen•10/31/24, 9:37 PM

Sec.

Isaac McFadyen

I

Isaac McFadyen•10/31/24, 9:38 PM

https://dash.cloudflare.com/?to=/:account/zone/caching/configuration

Isaac McFadyen

I

Isaac McFadyen•10/31/24, 9:38 PM

Try that

Isaac McFadyen

I

Isaac McFadyen•10/31/24, 9:38 PM

Then change Caching Level to "Ignore Query String"

Isaac McFadyen

I

Isaac McFadyen•10/31/24, 9:38 PM

Keep in mind that will ignore all query string not just the incremental cache bust there.

Isaac McFadyen

IIsaac McFadyen You could write a Pages Function to proxy some routes to another backend but tha...

Spunkie

S

Spunkie•10/31/24, 9:39 PM

Ya I thought about that but it kind of defeats why I wanted to use pages. Which is mostly to make it easy to understand and update for non technical users.

Spunkie

S

Spunkie•10/31/24, 9:39 PM

But that is prob what I'll do :thisisfine:

Isaac McFadyen

I

Isaac McFadyen•10/31/24, 9:39 PM

You could also create a Cache Rule that ignores a specific query string value but I believe it requires a Business plan

Mozzy

M

Mozzy•10/31/24, 9:39 PM

Something I failed to show is that I do need to bust cache if other search params change

Isaac McFadyen

I

Isaac McFadyen•10/31/24, 9:39 PM

Yeah one sec then

Mozzy

M

Mozzy•10/31/24, 9:39 PM

ahh

Isaac McFadyen

I

Isaac McFadyen•10/31/24, 9:40 PM

Unfortunately it requires Enterprise to exclude specific query params actually.

Mozzy

M

Mozzy•10/31/24, 9:41 PM

would a rewrite middleware that only keeps certain params be enough then?

Mozzy

M

Mozzy•10/31/24, 9:41 PM

Never done rewrites, so not sure it works

Isaac McFadyen

I

Isaac McFadyen•10/31/24, 9:42 PM

I believe that would work? Always hard to know because different kinds of Rules run in different orders and there's no real reference that I've seen

Isaac McFadyen

IIsaac McFadyen I believe that would work? Always hard to know because different kinds of Rules ...

Chaika

C

ChaikaOP•10/31/24, 9:43 PM

https://developers.cloudflare.com/ruleset-engine/reference/phases-list/#request-phases

Cloudflare Docs

Phases list | Cloudflare Ruleset Engine docs

The following tables list the phases of Cloudflare products powered by the Ruleset Engine, in the order those phases are executed. Some products such as the Cloudflare Web Application Firewall have more than one associated phase.

Chaika

CChaika https://developers.cloudflare.com/ruleset-engine/reference/phases-list/#request-...

Isaac McFadyen

I

Isaac McFadyen•10/31/24, 9:44 PM

Ah TY

Chaika

C

ChaikaOP•10/31/24, 9:44 PM

actual cache runs at the very end though, if the question is "can I use rules to modify what normal cache uses" pretty sure all of them work, transform headers at least def does

Flux

F

Flux•10/31/24, 10:17 PM

Wondering if anyone knows any ways to turn a PDF to a PNG/Image with workers. Is this even possible considering the v8 isolate constraints? (worst case setup lambda but dont want to)

Chaika

CChaika Then you have to use CF For SaaS, which doesn't support wildcards full stop for ...

elpupper

E

elpupper•10/31/24, 10:24 PM

ok so I'm doing some research and I dont think i get it. is CF for SaaS basically nginx? I cant use workers/worker routes here? Do I have to get my own server that cloudflare can point to?

Mozzy

MMozzy would a rewrite middleware that only keeps certain params be enough then?

Mozzy

M

Mozzy•10/31/24, 10:44 PM

This seems to do the trick, however cursed it is

export const stripUnusedSearchParams = createMiddleware<{ Bindings: CloudflareBindings }>(async (c, next) => {
  const url = new URL(c.req.url);
  let foundInvalidSearchParams = false;

  for (const key of url.searchParams.keys()) {
    if (!validSearchParams.includes(key)) {
      url.searchParams.delete(key);
      foundInvalidSearchParams = true;
    }
  }

  if (foundInvalidSearchParams) {
    return c.redirect(url.searchParams.size ? `${url.pathname}?${url.searchParams}` : url.pathname, 302);
  }

  await next();
});

export const stripUnusedSearchParams = createMiddleware<{ Bindings: CloudflareBindings }>(async (c, next) => {
  const url = new URL(c.req.url);
  let foundInvalidSearchParams = false;

  for (const key of url.searchParams.keys()) {
    if (!validSearchParams.includes(key)) {
      url.searchParams.delete(key);
      foundInvalidSearchParams = true;
    }
  }

  if (foundInvalidSearchParams) {
    return c.redirect(url.searchParams.size ? `${url.pathname}?${url.searchParams}` : url.pathname, 302);
  }

  await next();
});

Mozzy

M

Mozzy•10/31/24, 11:12 PM

sadly, req.url is readonly

Elliot Hesp

E

Elliot Hesp•11/1/24, 4:48 PM

Hey, if I have a RPC Target like so

class Foo extends RpcTarget {
  async fetch(request: Request): Promise<Response> {
    return new Response('Hello!');
  }
}

class Foo extends RpcTarget {
  async fetch(request: Request): Promise<Response> {
    return new Response('Hello!');
  }
}

Is there anyway I can just return that response from my worker?

At the moment, I get a type error:

Type '{ clone: Stub<() => Response>; status: number; statusText: string; headers: { get: Stub<(name: string) => string | null>; getAll: Stub<(name: string) => string[]>; ... 9 more ...; [Symbol.iterator]: Stub<...>; }; ... 12 more ...; blob: Stub<...>; } & Disposable' is not assignable to type 'Response'.
  The types of 'clone(...).status' are incompatible between these types.
    Type 'Promise<number>' is not assignable to type 'number'.ts(2322)

Type '{ clone: Stub<() => Response>; status: number; statusText: string; headers: { get: Stub<(name: string) => string | null>; getAll: Stub<(name: string) => string[]>; ... 9 more ...; [Symbol.iterator]: Stub<...>; }; ... 12 more ...; blob: Stub<...>; } & Disposable' is not assignable to type 'Response'.
  The types of 'clone(...).status' are incompatible between these types.
    Type 'Promise<number>' is not assignable to type 'number'.ts(2322)

bkyerv

B

bkyerv•11/1/24, 6:23 PM

is firebase sdk still not supported in workers?

Cross

C

Cross•11/1/24, 9:40 PM

Hi guys, i have a quick question, if i want to implement worker for my domain abc.com, on setting worker whats you think the best practice, set a route or a subdomain?

Geert Jan Sloos

G

Geert Jan Sloos•11/1/24, 11:04 PM

Is it difficult to create an Email Worker, and check if the incoming email has an PDF?

Omar Abid

O

Omar Abid•11/2/24, 2:21 PM

worker deployment is now failing 50% of the time for me. this has been on going for the last couple days.

Omar Abid

OOmar Abid worker deployment is now failing 50% of the time for me. this has been on going ...

elpupper

E

elpupper•11/2/24, 5:05 PM

using rust?

DoggeSlapper

D

DoggeSlapper•11/2/24, 5:22 PM

I wanna create an image with workers, using a base image + some text using specific fonts, can anyone recomend a library that works well on workers?

1984 Ford Laser

1

1984 Ford Laser•11/3/24, 1:22 AM

Does the

url

url

field of the draw object need to be a URL, or can it be an image object/blob? I have a second Worker that generates images, to be used as layers in an overlay, but I need to fetch the images to get dimensions before they're drawn as the overlay
https://developers.cloudflare.com/images/transform-images/draw-overlays/

Cloudflare Docs

Draw overlays and watermarks | Cloudflare Images docs

You can draw additional images on top of a resized image, with transparency and blending effects. This enables adding of watermarks, logos, signatures, vignettes, and other effects to resized images.

elpupper

Eelpupper using rust?

Omar Abid

O

Omar Abid•11/3/24, 3:23 AM

Both rust and javascript. it happens randomly.

Original message was deleted

fry69

F

fry69•11/4/24, 6:52 AM

See -> https://developers.cloudflare.com/workers/platform/limits/#daily-request

Type '{ clone: Stub<() => Response>; status: number; statusText: string; headers: { get: Stub<(name: string) => string | null>; getAll: Stub<(name: string) => string[]>; ... 9 more ...; [Symbol.iterator]: Stub<...>; }; ... 12 more ...; blob: Stub<...>; } & Disposable' is not assignable to type 'Response'.
  The types of 'clone(...).status' are incompatible between these types.
    Type 'Promise<number>' is not assignable to type 'number'.ts(2322)

Type '{ clone: Stub<() => Response>; status: number; statusText: string; headers: { get: Stub<(name: string) => string | null>; getAll: Stub<(name: string) => string[]>; ... 9 more ...; [Symbol.iterator]: Stub<...>; }; ... 12 more ...; blob: Stub<...>; } & Disposable' is not assignable to type 'Response'.
  The types of 'clone(...).status' are incompatible between these types.
    Type 'Promise<number>' is not assignable to type 'number'.ts(2322)

export const stripUnusedSearchParams = createMiddleware<{ Bindings: CloudflareBindings }>(async (c, next) => {
  const url = new URL(c.req.url);
  let foundInvalidSearchParams = false;

  for (const key of url.searchParams.keys()) {
    if (!validSearchParams.includes(key)) {
      url.searchParams.delete(key);
      foundInvalidSearchParams = true;
    }
  }

  if (foundInvalidSearchParams) {
    return c.redirect(url.searchParams.size ? `${url.pathname}?${url.searchParams}` : url.pathname, 302);
  }

  await next();
});

export const stripUnusedSearchParams = createMiddleware<{ Bindings: CloudflareBindings }>(async (c, next) => {
  const url = new URL(c.req.url);
  let foundInvalidSearchParams = false;

  for (const key of url.searchParams.keys()) {
    if (!validSearchParams.includes(key)) {
      url.searchParams.delete(key);
      foundInvalidSearchParams = true;
    }
  }

  if (foundInvalidSearchParams) {
    return c.redirect(url.searchParams.size ? `${url.pathname}?${url.searchParams}` : url.pathname, 302);
  }

  await next();
});