You can fill out the increase request form @ to try and request increases

But a plan based one is probably ent only

Yeahhh probably

Also, does Date.now() update after doing a network request? I thought I saw that somewhere

It does after any I/O

Will the following appear when using this quick edit

I don't know what that says, sorry

sorry

I believe it works fine on the new one, yep

okay, thank you

I'm following the documentation on the workers site for getting headers.

const headers = new Headers();
    console.log(headers.get('name'))

It keeps coming back null. I'm sending a custom here with the request from postman

Not sure what you're expecting that code snippet to do - you make an empty headers object, then check if it has something in it 😅

Did you mean

request.headers.get('name')

?

You probably want

req.headers.get('name')

? You're creating an empty

Headers

object there and then reading from it... yeah what Skye said

Would I add the request part in the

addEventListener('fetch', (e) => {
    e.respondWith(router.handle(e.request))
});

or the

router.get("/", () => {
    return new Response("Hello World")
});

It looks like you're passing the

request

to your router, so you'd need to modify your router handler function like this:

js
router.get("/", (request) => {
    const name = request.headers.get('name'); // here
    return new Response("Hello World")
});

Note the

request

argument in the handler there primarily

Is a good pricing estimator anywhere for websocket workers + durable objects?

You can sort of use https://pricing.ceru.dev - every incoming connection and incoming websocket message counts as a request, and any time a DO instance has a ws connected to it, counts as runtime

Thanks! Man, as a guess it looks like long-running high-interactivity tabletop games might be too expensive to be feasible in Workers 😦 I have a card game where you drop cards in a turnless fashion so clients needs to be in a connected state to get updates in a resonable time. A tabletop runs continuously, we'd need at least one running 24/7, and an ideal state would be running hundreds or thousands of tables at the same time. A single $20 VM instance could probably handle hundreds as a guess, but it sounds like it'd be orders of magnitude more pricey with workers.

Which sucks 'cause Workers does seem like the right solution for every other problem I'm trying to solve other than cost.

% agree, there is something in the works that means inactive DOs won't get charged

Thank you that worked

How are IPs assigned to workers? For example if worker A makes an HTTP request to an external server, and worker B makes an HTTP request to the same server, will the server view them as having the same IP, different ones, or are these rotated somehow?

In that example they'll be Cloudflare egress IPs (https://www.cloudflare.com/ips/)

if you were to fetch another Cloudflare zone, it'd be a static IP identifying Workers

In my case they are in the same zone. So each worker will get assigned a different egress IP?

It’s not related to the Worker

It’s the same as if you just had the website proxied