Then why would you ever need a scheduled worker to do the updates ?

You make a very good point lol

js
let cacheKey = `https://${url.host}${url.pathname}`;
let data = await caches.default.match(cacheKey);

if (data) {
    data = await data.json();
    return Response.json({ data: data.data, });
}


let catResponse = await fetch('https://cataas.com/cat?json=true');
catResponse = await catResponse.json();

let cachedAPILayer = Response.json({ data: catResponse });

cachedAPILayer.headers.set("Cache-Control", `s-maxage=5`);
ctx.waitUntil(caches.default.put(cacheKey, cachedAPILayer.clone()));

return cachedAPILayer;

caches random api i found for 5 sec @ https://test.svge.ms/cache

lots of reused code so variables are not clean lol

Appreciated! I'll sleep on it and figure something out

Thank you both for your help 🙏

s-maxage=5 hmm

You can make it longer if you need, maybe make it 86400 if you want. And you can choose to cache upstream response directly, or to cache "your" own generated response. You can cache the response right before you return.

Cache API has 2 bad parts:

1. A cache entry may be deleted before expiration for no reason.

2. Cache operations are count as outgoing sub requests

Hey! if i make a subdomain. abc.mydomain, and from workers i want to fetch abc.mydomain:1234, will it work workers will be for example on xyz.mydomain

It will go direct to origin, provided the Worker runs on your domain and, in the case of Custom Domains (and not Routes), runs on the specific subdomain.

so it will work?

oh, specific subdomain?

If the Worker is setup with a Custom Domain, it needs to be the on the actual domain as far as I understand it.

what is a zone @Skye

is it a domain?

or subdomain

I believe in this case subdomain

😢

If it's set-up as a custom domain, yes.

Is it possible to run Puppeteer or Playwright inside a worker using something like Browserless? Like does it run? Or do those packages use node api's regardless?

Coming : https://blog.cloudflare.com/introducing-workers-browser-rendering-api/

I saw that and that is very cool! But for now could I do this in theory with Browserless?

No. AFAIK,

browserless

relies on Headless Chromium, which while it might work in Workers(if compiled to WebAssembly), has the distinct disadvantage of being too large for the 1-5 MB limit for a deployed script.

What are you trying to render, if I may ask?

Browserless runs the browser in their environment, right? And I would use the playwright package to connect to that browser? Not sure, that's why I came here to ask 😅

I'm trying to render social images for now, maybe more stuff later