you still don't support any method of knowing if an email was ever delivered, right?

Sskoogeer The API will continue working. We have been providing email delivery service con...

J

James•6/21/24, 9:33 PM

For how long will the API key you provide to folks who DM you continue to work? Is there a timeline to when you might require them to pay for this "commercial service"?

J

James•6/21/24, 9:33 PM

This is not clear at all in your support article right now

Ddave you still don't support any method of knowing if an email was ever delivered, ri...

S

skoogeer•6/21/24, 9:34 PM

Hi Dave. If you sign up for an API key, you will get a console account that lets you see logs of message deliveries. There is not webhook for delivery status yet, but this is a key feature that we will provide when we relaunch later this year (among many more).

D

daveOP•6/21/24, 9:34 PM

btw AWS SES works perfectly with aws4fetch, we've been using it for 100% of our outgoing company emails

Sskoogeer Hi Dave. If you sign up for an API key, you will get a console account that lets...

D

daveOP•6/21/24, 9:35 PM

uh, so I have to fill this form out? https://info.mailchannels.com/contact-sales

Contact Sales

Contact MailChannels to learn more about how we can help secure and deliver your hosting emails.

JJames This is not clear at all in your support article right now

S

skoogeer•6/21/24, 9:41 PM

Thanks for the feedback, James.

Ddave uh, so I have to fill this form out? https://info.mailchannels.com/contact-sales

S

skoogeer•6/21/24, 9:41 PM

Yes, at this time, the general sales contact form is the way to get an API key. We will issue keys and accounts to everyone next week.

Sskoogeer Thanks for the feedback, James.

J

James•6/21/24, 9:42 PM

Can you provide a clear answer please? If you're going to be soliciting folks to DM you here, you need to be extremely clear and transparent about what you're offering, and when/if that will change. Your communication up to this point hasn't been the greatest, so you need to go an extra mile here if you want to create any goodwill in the developer community.

Sskoogeer Yes, at this time, the general sales contact form is the way to get an API key. ...

D

daveOP•6/21/24, 9:44 PM

could you give us a copy/paste message to send to support?

S

skoogeer•6/21/24, 9:48 PM

Sure, and apologies for the lack of clarity. When you contact our team, just indicate you are a Cloudflare Workers customer and you would like an API key. The form collects your name and email address so that we can follow up with the necessary credentials. My team is on the lookout for these submissions.

JJames Can you provide a clear answer please? If you're going to be soliciting folks to...

S

skoogeer•6/21/24, 9:56 PM

I appreciate your request for a specific timeline and recommend that you sign up for a free API key as we plan to engage actively with users once they have accounts on our platform.

E

euan•6/21/24, 9:58 PM

Sounds like a fine opportunity for Cloudflare to introduce a platform-native email sending product for Workers to leverage. Shame it's not developer week.

L

louis.ecxs•6/21/24, 10:00 PM

uh, hi everyone i’m really not knowledgeable on this stuff but i got the advanced certificate and a bunch of subdomains, but one of my workers, that i deployed with my Debian Raspberrypi with Vue3 and Vite is now pointing all of my subdomains to “fallback domain” which stupidly enough, is - by my own admission- my root domain.

now i’ve changed the fallback but cloudflare hasn’t validated the root cert in 48h, while all subdomains are instantly ready.
so i guess my question is, if my worker is removed, does it delete all subs?

i’m ashamed of it but that’s the worker and i’m able to configure it but idk why and when i made it

Sskoogeer I appreciate your request for a specific timeline and recommend that you sign up...

J

James•6/21/24, 10:01 PM

To be clear, you want people to signup, hand over their details, use your platform, build out their implementations, and then you'll tell them when they might no longer be able to use it, pay up, or functionality will change? This isn't a good look. You're smarter than this, come on.

J

James•6/21/24, 10:04 PM

I appreciate you coming here and offering this to folks, but you really need to be more explicit about what you're offering. You once offered a free service with Cloudflare Workers that was then terminated with very little notice. If you want people to trust you and your platform again, you can't not provide them with more explicit information and guarantees Ken.

Llouis.ecxs uh, hi everyone i’m really not knowledgeable on this stuff but i got the advance...

L

louis.ecxs•6/21/24, 10:05 PM

sorry - that’s the worker:

// src/index.ts
var src_default = {
async fetch(request, env, ctx) {
return new Response("Hello World!");
}
};
export {
src_default as default
};
//# sourceMappingURL=index.js.map

Sskoogeer Sure, and apologies for the lack of clarity. When you contact our team, just ind...

D

daveOP•6/21/24, 10:06 PM

thanks, mind pinging me when you have something ready I can copy paste for support?

S

skoogeer•6/21/24, 10:13 PM

I understand your concerns, and I want to clarify a few points:

Using our service is completely optional. If the new API key requirement doesn't work for you, you're free to use an alternative.
The API will remain free. The key simply helps us manage usage and ensure sustainability.
We aim to be transparent, but can't always pre-announce changes publicly for competitive reasons. However, we commit to giving registered users ample notice of any major changes impacting their implementation.
Signing up for a key now won't impact your current functionality. If any changes are necessary in the future, we'll work with you to ensure a smooth transition.

Our goal is to balance providing a useful free service with keeping our business viable. API keys are a step towards that. I appreciate your feedback as we navigate this change. Please let me know if you have any other concerns.

Ddave thanks, mind pinging me when you have something ready I can copy paste for suppo...

S

skoogeer•6/21/24, 10:14 PM

Hi Dave. Sorry, I thought I pasted something above. Just indicate you are a Cloudflare Workers customer and you would like an API key. The form collects your name and email address so that we can follow up with the necessary credentials. My team is on the lookout for these submissions.

Sskoogeer I understand your concerns, and I want to clarify a few points: 1. Using our se...

J

James•6/21/24, 10:16 PM

To confirm, with this quote in your article:

In exchange for being an early user, you will receive a significantly discounted price when the improved commercial service launches later this year.

This new "commercial service" won't replace the free service, and folks will be able to continue using that if they get an API key?

J

James•6/21/24, 10:17 PM

I don't think anyone here is against getting an API key - that's standard practice. There's just lots of concerns about reliability, deprecation, etc. based on previous experience and communication here.

Sskoogeer Hi Dave. Sorry, I thought I pasted something above. Just indicate you are a Clou...

D

daveOP•6/21/24, 10:18 PM

alright, I sent

Cloudflare Workers MailChannels developer/user, we'd like a free API key please.

Cloudflare Workers MailChannels developer/user, we'd like a free API key please.

Cloudflare Workers MailChannels developer/user, we'd like a free API key please.

Cloudflare Workers MailChannels developer/user, we'd like a free API key please..

JJames To confirm, with this quote in your article: > In exchange for being an early us...

D

daveOP•6/21/24, 10:18 PM

oh wait, just discounted?! I thought there was gonna be a free tier

Ddave oh wait, just discounted?! I thought there was gonna be a free tier

J

James•6/21/24, 10:19 PM

that's where the confusion is and still hasn't really been clarified. The first paragraph in https://support.mailchannels.com/hc/en-us/articles/26814255454093-End-of-Life-Notice-Cloudflare-Workers seems to contradict itself.

JJames that's where the confusion is and still hasn't really been clarified. The first ...

D

daveOP•6/21/24, 10:24 PM

9 days of notice is crazy either way imo

Ddave 9 days of notice is crazy either way imo

J

James•6/21/24, 10:25 PM

yeah I would imagine most people have migrated away at this point, as per advice of lots of folks in this server and the community

JJames yeah I would imagine most people have migrated away at this point, as per advice...

D

daveOP•6/21/24, 10:26 PM

Even if they extend it to 9 years, it's a super bad idea to keep doing business with a company that thought 9 days of notice would be okay to begin with.

J

James•6/21/24, 10:26 PM

Also another question @skoogeer after you've covered the previous, in that article you include:

MailChannels plans to launch a more capable email delivery API in the future, complete with delivery tracking features, DKIM-key hosting, generative-AI delivery optimization, and more.

Is this going to include basic domain sender verification this time?

JJames Also another question @skoogeer after you've covered the previous, in that artic...

D

daveOP•6/21/24, 10:36 PM

that's a feature to not have! ;P

D

daveOP•6/21/24, 11:07 PM

do I need to set

version_metadata

version_metadata

version_metadata

version_metadata for each of my envs, or is the key enough to have in the root just once?

D

daveOP•6/21/24, 11:09 PM

answer, I need it for each env

D

daveOP•6/21/24, 11:17 PM

@Unsmart general-discussions

I found a good solution for this (Backblaze B2 erroring with 503, 429, 500, etc):

Setup: I set up two buckets, one in us-east and one in us-west. Then, I set up two bidirectional cloud replication rules.

In my Worker, I try to upload to the us-east bucket first, and flip flop between us-west and us-east for repeated attempts until it works.

Eventually, both buckets will contain 1:1 of the same data, even when there's an outage too in one region.

K

kindaraffy•6/22/24, 10:24 AM

Hi, you guys might want to extend the cf docs on worker RPC to include the utility type

InstanceType

InstanceType

InstanceType

InstanceType. It automatically types your RPC classes so you could use your RPC methods in other workers with type-safety.

https://developers.cloudflare.com/workers/runtime-apis/rpc/

An example:

import { RpcTarget, WorkerEntrypoint } from 'cloudflare:workers';

class Counter extends RpcTarget {
    #value = 0;

    increment(amount: number = 1) {
        this.#value += amount;
        return this.#value;
    }

    get value() {
        return this.#value;
    }
}

export class CounterService extends WorkerEntrypoint {
    async newCounter() {
        return new Counter();
    }
}

export type CounterServiceStub = InstanceType<typeof CounterService>;

import { RpcTarget, WorkerEntrypoint } from 'cloudflare:workers';

class Counter extends RpcTarget {
    #value = 0;

    increment(amount: number = 1) {
        this.#value += amount;
        return this.#value;
    }

    get value() {
        return this.#value;
    }
}

export class CounterService extends WorkerEntrypoint {
    async newCounter() {
        return new Counter();
    }
}

export type CounterServiceStub = InstanceType<typeof CounterService>;

import { RpcTarget, WorkerEntrypoint } from 'cloudflare:workers';

class Counter extends RpcTarget {
    #value = 0;

    increment(amount: number = 1) {
        this.#value += amount;
        return this.#value;
    }

    get value() {
        return this.#value;
    }
}

export class CounterService extends WorkerEntrypoint {
    async newCounter() {
        return new Counter();
    }
}

export type CounterServiceStub = InstanceType<typeof CounterService>;

import { RpcTarget, WorkerEntrypoint } from 'cloudflare:workers';

class Counter extends RpcTarget {
    #value = 0;

    increment(amount: number = 1) {
        this.#value += amount;
        return this.#value;
    }

    get value() {
        return this.#value;
    }
}

export class CounterService extends WorkerEntrypoint {
    async newCounter() {
        return new Counter();
    }
}

export type CounterServiceStub = InstanceType<typeof CounterService>;

Then, use the CounterServiceStub in your RPC client for type-checked method signatures for the downstream worker you are calling.

Cloudflare Docs

Remote-procedure call (RPC) · Cloudflare Workers docs

The built-in, JavaScript-native RPC system built into Workers and Durable Objects.

Kkindaraffy Hi, you guys might want to extend the cf docs on worker RPC to include the utili...

K

kian•6/22/24, 6:28 PM

That is what the

Service

Service

Service

Service and

DurableObjectNamespace

DurableObjectNamespace

DurableObjectNamespace

DurableObjectNamespace generic types do, isn't it?

https://developers.cloudflare.com/workers/runtime-apis/rpc/typescript/

J

johtso•6/22/24, 6:33 PM

What's up with not being able to deploy workers with

logpush = true

logpush = true

logpush = true

logpush = true when you're a member of another account? https://github.com/cloudflare/workers-sdk/issues/2991

J

johtso•6/22/24, 6:33 PM

isn't this an issue that loads of people should be bumping into? i'm confused as to why the issue was closed

J

johtso•6/22/24, 6:33 PM

not sure how removing logpush from your config is a solution

K

kian•6/22/24, 6:37 PM

If you're having issues with Logpush causing API errors then I'd recommend opening a new issue and just reference that one since it's more than a year old. Unless your account isn't able to use Logpush, it shouldn't cause errors.

H

Harris•6/22/24, 6:39 PM

Has anyone encountered a situation where trying to listen to the live log stream of a durable object, just causes all websocket connections to fail?

I can't have my game clients connect to the durable object and listen to live logs at the same time and it makes debugging quite hard

H

Harris•6/22/24, 6:39 PM

I get

{
      "name": "Error",
      "message": "Subrequest depth limit exceeded. This request recursed through Workers too many times. This can happen e.g. if you have a Worker or Durable Object that calls other Workers or objects recursively.",
      "timestamp": 1719081484425
    }

{
      "name": "Error",
      "message": "Subrequest depth limit exceeded. This request recursed through Workers too many times. This can happen e.g. if you have a Worker or Durable Object that calls other Workers or objects recursively.",
      "timestamp": 1719081484425
    }

{
      "name": "Error",
      "message": "Subrequest depth limit exceeded. This request recursed through Workers too many times. This can happen e.g. if you have a Worker or Durable Object that calls other Workers or objects recursively.",
      "timestamp": 1719081484425
    }

{
      "name": "Error",
      "message": "Subrequest depth limit exceeded. This request recursed through Workers too many times. This can happen e.g. if you have a Worker or Durable Object that calls other Workers or objects recursively.",
      "timestamp": 1719081484425
    }

no matter what I do, but only when I try to listen to live logs

edit: https://community.cloudflare.com/t/worker-throwing-recursion-error-when-calling-a-durable-object-even-though-there-is/640749/4 i see, is an error with smart mode

Jjohtso What's up with not being able to deploy workers with `logpush = true` when you'r...

W

Walshy•6/22/24, 6:42 PM

What's the error you hit?

WWalshy What's the error you hit?

J

johtso•6/22/24, 6:42 PM

workers.api.error.unauthorized [code: 10023]

W

Walshy•6/22/24, 6:43 PM

Sounds like that account doesn't have workers paid

J

johtso•6/22/24, 6:43 PM

ooooo

J

johtso•6/22/24, 6:43 PM

J

johtso•6/22/24, 6:43 PM

friendly error messages would be pretty awesome

W

Walshy•6/22/24, 6:44 PM

Workers errors are not always the clearest today :blobsweats:
cc: @tanushree │ Workers PM it'd be good to do an error audit at some point

J

johtso•6/22/24, 7:05 PM

oof, "Can't read from request stream after response has been sent." only seemed to happen when deployed, not when running locally

J

johtso•6/22/24, 7:05 PM

I guess it's a wrangler inconsistency

J

johtso•6/22/24, 9:25 PM

does this look right for forking a request to two separate endpoints?

  const [body, body2] = request.body.tee()
  const result = await Promise.allSettled([
    fetch(url1, { method: request.method, headers: request.headers, body: body }),
    fetch(url2, { method: request.method, headers: request.headers, body: body2 })
  ])

  const [body, body2] = request.body.tee()
  const result = await Promise.allSettled([
    fetch(url1, { method: request.method, headers: request.headers, body: body }),
    fetch(url2, { method: request.method, headers: request.headers, body: body2 })
  ])

  const [body, body2] = request.body.tee()
  const result = await Promise.allSettled([
    fetch(url1, { method: request.method, headers: request.headers, body: body }),
    fetch(url2, { method: request.method, headers: request.headers, body: body2 })
  ])

  const [body, body2] = request.body.tee()
  const result = await Promise.allSettled([
    fetch(url1, { method: request.method, headers: request.headers, body: body }),
    fetch(url2, { method: request.method, headers: request.headers, body: body2 })
  ])

you still don't support any method of knowing if an email was ever delivered, right?

Similar Threads

Similar Threads

Similar Threads