i like to have things seperated, because Astro is just a simple PoC but the backend is already bigge

F

F0rceOP•2/23/24, 4:06 PM

and with two seperate "projects" I am not limited to a specific ui framework

FF0rce i like to have things seperated, because Astro is just a simple PoC but the back...

H

Hello, I’m Allie!•2/23/24, 4:07 PM

You can isolate the backend into a different package, then have a simple shim within Astro that passes the request into the backend

H

Hello, I’m Allie!•2/23/24, 4:07 PM

Then you can keep them separate, but within the same app

H

Hello, I’m Allie!•2/23/24, 4:07 PM

And the backend is always in sync with the frontend

F

F0rceOP•2/23/24, 4:09 PM

im already using rfc (from hono) to have end to end safety, so its always in sync. but Ive never worked with a shim.

F

F0rceOP•2/23/24, 4:10 PM

i was just curios if this setup works correcly, thanks for the input

F

F0rceOP•2/23/24, 4:11 PM

ahhh another "bonus point" with my current setup is that its easier to interact with Cloudflare Bindings (especially D1). I really did not enjoy how astro handles bindings...

F

F0rceOP•2/23/24, 4:11 PM

and Astro does not support service bindings in the

@astrojs/cloudflare

@astrojs/cloudflare

@astrojs/cloudflare

@astrojs/cloudflare adapter yet..

FF0rce ahhh another "bonus point" with my current setup is that its easier to interact ...

H

Hello, I’m Allie!•2/23/24, 4:13 PM

If you are willing to get a little jiggy, you could probably make your “Worker” call Astro instead?

H

Hello, I’m Allie!•2/23/24, 4:13 PM

But yeah, this seems more complex

B

bun•2/23/24, 4:13 PM

is the

environments

environments

for pages same as for workers? meaning this: https://developers.cloudflare.com/workers/wrangler/environments/

F

F0rceOP•2/23/24, 4:14 PM

just out of curiosity, when I call

fetch(/api)

fetch(/api)

in Astro this will probably not work because the request will not go through the edge network (completely), right ? - I will have to use the full url... or am I mistaken

Bbun is the `environments` for pages same as for workers? meaning this: https://devel...

H

Hello, I’m Allie!•2/23/24, 4:15 PM

Not really. Pages uses branches, though you can only define explicit environments for “production”, and “everything else”

B

bun•2/23/24, 4:15 PM

how do i configure different d1 databases in my wrangler.toml for my pages

B

bun•2/23/24, 4:15 PM

thats all im trying tp do

FF0rce just out of curiosity, when I call `fetch(/api)` in Astro this will probably not...

H

Hello, I’m Allie!•2/23/24, 4:15 PM

Not sure, actually. Might be a case of “try it and find out”

Bbun how do i configure different d1 databases in my wrangler.toml for my pages

H

Hello, I’m Allie!•2/23/24, 4:15 PM

For dev, or when deployed?

B

bun•2/23/24, 4:16 PM

both

B

bun•2/23/24, 4:16 PM

i believe for local it only uses a local sqlite database for some reason (since d1 cant be used locally)

B

bun•2/23/24, 4:16 PM

im on

next-on-pages

next-on-pages

HHello, I’m Allie!Not sure, actually. Might be a case of “try it and find out”

F

F0rceOP•2/23/24, 4:16 PM

If im not mistaken I already tried that out a bit ago, and got 404 all over the board... then I had to deploy my worker under a new subdomain (like api.dashboard.mydomain.com)... but Astro (and wrangler) received a lot of updates... xd

F

F0rceOP•2/23/24, 4:18 PM

but I really hate that setup xd. I like isolation (per process / page).

Bbun both

H

Hello, I’m Allie!•2/23/24, 4:18 PM

You can use the wrangler.toml for local, but I don’t think that supports environments. As for the deployed version, that has to be configured via the dash

B

bun•2/23/24, 4:18 PM

lel

FF0rce If im not mistaken I already tried that out a bit ago, and got 404 all over the ...

H

Hello, I’m Allie!•2/23/24, 4:19 PM

Did you deploy it as a route, or as a Custom Domain?

B

bun•2/23/24, 4:19 PM

well i guess i can pass in

--d1

--d1

arguments

--kv

--kv

arguments manually

HHello, I’m Allie!Did you deploy it as a route, or as a Custom Domain?

F

F0rceOP•2/23/24, 4:19 PM

first as a route really like dashboard.mydomain.com/api and secondly as custom domain api.dashboard.mydomain.com

The astro page is always under custom domain (dashboard.mydomain.com)

Bbun well i guess i can pass in `--d1` arguments `--kv` arguments manually

H

Hello, I’m Allie!•2/23/24, 4:19 PM

That’s probably the easiest, though it looks kind of ugly

H

Hello, I’m Allie!•2/23/24, 4:20 PM

Put it in your package.json scripts so you don’t have to remember

B

bun•2/23/24, 4:20 PM

im eating spaghetti with mushrooms right now its delicious

B

bun•2/23/24, 4:20 PM

oops.. wrong dm

FF0rce first as a route really like dashboard.mydomain.com/api and secondly as custom d...

H

Hello, I’m Allie!•2/23/24, 4:20 PM

Do you see requests in the tail logs for the Worker?

Bbun im eating spaghetti with mushrooms right now its delicious

F

F0rceOP•2/23/24, 4:20 PM

im jealous xd

Bbun im eating spaghetti with mushrooms right now its delicious

H

Hello, I’m Allie!•2/23/24, 4:21 PM

Damn, making me hungry now…

HHello, I’m Allie!Do you see requests in the tail logs for the Worker?

F

F0rceOP•2/23/24, 4:21 PM

as I said, I tried that really long ago (proably 8+ months ago) and just wanted to ask before I go all in in that direction aswell. I can come back if I did some tests

HHello, I’m Allie!Damn, making me hungry now…

F

F0rceOP•2/23/24, 7:31 PM

sorry for pinging you / answering. But ive just spent time deploying everything (or making a good staging environment, with Zero Trust, etc -- the basics). But a request from dashboard.mydomain.com/ping to dashboard.mydomain.com/api/v1/ping does not work (it just returns 404 and the tail does not see the request registered). I dug a bit deeper and do I have to exclude / include a specific route in the _routes.json that is generated by astro?

Left backend (worker) right frontend (pages)

Bildschirmfoto_2024-02-23_um_20.31.16.png

F

F0rceOP•2/23/24, 7:32 PM

is there a way to explicity tell the page that some routes should NOT be handled by the page itself because something else is taking care of it (in that case the worker) ?

F

F0rceOP•2/23/24, 7:34 PM

when using

exlude: ["/api/*"]

exlude: ["/api/*"]

I get the following error from pages:

(log) TypeError: Fetch API cannot load: /api/v1/ping

(log) TypeError: Fetch API cannot load: /api/v1/ping

I'll try with absolute url

FF0rce sorry for pinging you / answering. But ive just spent time deploying everything ...

H

Hello, I’m Allie!•2/23/24, 7:38 PM

If the Worker is being called at all, then it should be called before Pages

F

F0rceOP•2/23/24, 7:38 PM

in the screenshot you can see that it is not called at all, because the pages tail has the api/v1/ping call

FF0rce in the screenshot you can see that it is not called at all, because the pages ta...

H

Hello, I’m Allie!•2/23/24, 7:39 PM

Yeah, so something else is happening. What is the route for your Worker?

F

F0rceOP•2/23/24, 7:39 PM

route = { pattern = "dashboard.ruxiom.dev/api/*", zone_name = "ruxiom.dev" }

route = { pattern = "dashboard.ruxiom.dev/api/*", zone_name = "ruxiom.dev" }

route = { pattern = "dashboard.ruxiom.dev/api/*", zone_name = "ruxiom.dev" }

route = { pattern = "dashboard.ruxiom.dev/api/*", zone_name = "ruxiom.dev" }

F

F0rceOP•2/23/24, 7:41 PM

and the Page has a custom domain pointing to

dashboard.ruxiom.dev

dashboard.ruxiom.dev

F

F0rceOP•2/23/24, 7:47 PM

absolute url with excluding does not work...

S

Simciutoja•2/23/24, 7:53 PM

0:23:57.383    Deploying your site to Cloudflare's global network...
20:24:00.014    Uploading... (0/26)
20:24:00.742    Uploading... (9/26)
20:24:00.898    Uploading... (17/26)
20:24:01.763    Uploading... (26/26)
20:24:01.763    ✨ Success! Uploaded 26 files (2.31 sec)
20:24:01.763    
20:24:02.295    ✨ Upload complete!
20:24:04.409    Success: Assets published!
20:24:05.495    Success: Your site was deployed!

0:23:57.383    Deploying your site to Cloudflare's global network...
20:24:00.014    Uploading... (0/26)
20:24:00.742    Uploading... (9/26)
20:24:00.898    Uploading... (17/26)
20:24:01.763    Uploading... (26/26)
20:24:01.763    ✨ Success! Uploaded 26 files (2.31 sec)
20:24:01.763    
20:24:02.295    ✨ Upload complete!
20:24:04.409    Success: Assets published!
20:24:05.495    Success: Your site was deployed!

0:23:57.383    Deploying your site to Cloudflare's global network...
20:24:00.014    Uploading... (0/26)
20:24:00.742    Uploading... (9/26)
20:24:00.898    Uploading... (17/26)
20:24:01.763    Uploading... (26/26)
20:24:01.763    ✨ Success! Uploaded 26 files (2.31 sec)
20:24:01.763    
20:24:02.295    ✨ Upload complete!
20:24:04.409    Success: Assets published!
20:24:05.495    Success: Your site was deployed!

0:23:57.383    Deploying your site to Cloudflare's global network...
20:24:00.014    Uploading... (0/26)
20:24:00.742    Uploading... (9/26)
20:24:00.898    Uploading... (17/26)
20:24:01.763    Uploading... (26/26)
20:24:01.763    ✨ Success! Uploaded 26 files (2.31 sec)
20:24:01.763    
20:24:02.295    ✨ Upload complete!
20:24:04.409    Success: Assets published!
20:24:05.495    Success: Your site was deployed!

S

Simciutoja•2/23/24, 7:54 PM

could someone help? I'v trying to upload strapi.

S

Simciutoja•2/23/24, 7:54 PM

on localhost its works but unfortunetly there not.

F

F0rceOP•2/23/24, 8:14 PM

the wierd thing is that a normal request (from bruno, local, or the browser itself) succeedes. only the sort of "subrequest" to an worker route from a cloudflare page is not working as expected (or by design)

F

F0rceOP•2/23/24, 9:00 PM

ok I was able to finally sort it out. But this seems like a super hacky workaround, there has to be a way to do without my current doings..... (worker sites had the exact use-case I am looking for right now).

Basically in Astro I created an

/api/[...request].js

/api/[...request].js

route (which is an api route) in my

pages

pages

folder.
The content is pretty simple:

export async function GET({ params }) {
  const backendRequest = await fetch(
    "https://dashboard.ruxiom.dev/api/" + params.request,
    {
      headers: {
        "CF-Access-Client-Id": "something.access",
        "CF-Access-Client-Secret":
          "something different",
      },
    }
  );

  return backendRequest.clone();
}

export async function GET({ params }) {
  const backendRequest = await fetch(
    "https://dashboard.ruxiom.dev/api/" + params.request,
    {
      headers: {
        "CF-Access-Client-Id": "something.access",
        "CF-Access-Client-Secret":
          "something different",
      },
    }
  );

  return backendRequest.clone();
}

export async function GET({ params }) {
  const backendRequest = await fetch(
    "https://dashboard.ruxiom.dev/api/" + params.request,
    {
      headers: {
        "CF-Access-Client-Id": "something.access",
        "CF-Access-Client-Secret":
          "something different",
      },
    }
  );

  return backendRequest.clone();
}

export async function GET({ params }) {
  const backendRequest = await fetch(
    "https://dashboard.ruxiom.dev/api/" + params.request,
    {
      headers: {
        "CF-Access-Client-Id": "something.access",
        "CF-Access-Client-Secret":
          "something different",
      },
    }
  );

  return backendRequest.clone();
}

As you can see I had to add an additional Service Auth Include Rule in our Zero Trust Dashboard, without it I get the error that I am not allowed to view that endpoint.

This seems absolutely ridicoulus, it really does...

F

F0rceOP•2/23/24, 9:00 PM

cc @Hello, I’m Allie!

i like to have things seperated, because Astro is just a simple PoC but the backend is already bigge

Similar Threads

Similar Threads

Similar Threads