Question: if a worker exceeds its 6 simultaneous network connection count, is it the case that it wi

O

ogshadowfiendOP•4/11/24, 7:11 PM

Ahhh.

O

ogshadowfiendOP•4/11/24, 7:11 PM

Hmm okay will take a closer look here.

O

ogshadowfiendOP•4/11/24, 7:12 PM

Yeah, 500ms I was looking at a different thing. It still looks like we're closer to 20-25s than 30s.

O

ogshadowfiendOP•4/11/24, 7:13 PM

But actually that's wall clock time, the dashboard is saying 500ms CPU time:

O

ogshadowfiendOP•4/11/24, 7:13 PM

O

ogshadowfiendOP•4/11/24, 7:14 PM

That seems like the next possible culprit but hard to figure out how we'd get there, this is a pretty lightweight process most of whose work is on the network.

O

ogshadowfiendOP•4/11/24, 7:14 PM

Will dig a little more though, maybe there's something we're missing.

O

ogshadowfiendOP•4/11/24, 7:15 PM

Thank you!

A

alotabits•4/11/24, 7:16 PM

Do conditional PutObject requests to R2 count as writes, for billing, if the conditon fails?

Oogshadowfiend Question: if a worker exceeds its 6 simultaneous network connection count, is it...

O

ogshadowfiendOP•4/11/24, 7:16 PM

Still interested in the answer to this question though: if we push the past 6 connections limit, do we get a different outcome or is it still reported as

exceededCpu

exceededCpu

?

Oogshadowfiend Still interested in the answer to this question though: if we push the past 6 co...

C

Chaika•4/11/24, 7:18 PM

no, it'd just cancel the oldest and if your code wasn't setup to check for it, you'd script error

C

Chaika•4/11/24, 7:18 PM

If the system detects that a Worker is deadlocked on open connections — for example, if the Worker has pending connection attempts but has no in-progress reads or writes on the connections that it already has open — then the least-recently-used open connection will be canceled to unblock the Worker.

If the Worker later attempts to use a canceled connection, an exception will be thrown. These exceptions should rarely occur in practice, though, since it is uncommon for a Worker to open a connection that it does not have an immediate use for.

https://developers.cloudflare.com/workers/platform/limits/#simultaneous-open-connections

O

ogshadowfiendOP•4/11/24, 7:21 PM

Got it. Ok. Well I think my hypothetical fix for network connections will also potentially help with memory usage so let's see how it goes haha.

O

ogshadowfiendOP•4/11/24, 8:11 PM

(It went just fine.)

J

johtso•4/11/24, 11:31 PM

How do I stop my worker from connecting to live services when running locally, and instead wait for me to run a local instance of the service?

Jjohtso How do I stop my worker from connecting to live services when running locally, a...

D

dave•4/12/24, 12:23 AM

is the local instance of that other service also another Worker, or something that's just HTTP?

J

johtso•4/12/24, 12:24 AM

Another worker over rpc

J

johtso•4/12/24, 12:24 AM

I get a warning from wrangler that it's bound to a live service

C

codebam•4/12/24, 12:57 AM

when I enable warp vpn I can't wrangler tail

ZZiga Zajc Click to see attachment

R

rootalien•4/12/24, 2:12 AM

From there you can select the Respect cache option.

G

George Boot•4/12/24, 9:47 AM

If my worker is only used for egressing objects from R2, does it make a functional differnce to use the r2 binding vs the s3 api from r2? We create new buckets daily and want to get rid of us having to reploy the workers to add the new bucket binding

GGeorge Boot If my worker is only used for egressing objects from R2, does it make a function...

H

Hello, I’m Allie!•4/12/24, 9:56 AM

Bindings are generally faster than the API, due to not running over HTTP

H

Hello, I’m Allie!•4/12/24, 9:57 AM

Not to say you can't use the S3 API, just that the difference might be noticeable

HHello, I’m Allie!Not to say you can't use the S3 API, just that the difference might be noticeabl...

G

George Boot•4/12/24, 9:58 AM

thanks! Will there ever be something to bind to a group or to all workers?

GGeorge Boot thanks! Will there ever be something to bind to a group or to all workers?

H

Hello, I’m Allie!•4/12/24, 9:58 AM

Dynamic Bindings are something they are working on, but not sure if that specifically applies to R2...

G

George Boot•4/12/24, 9:59 AM

But great, thanks for the info!

V

Victor•4/12/24, 10:19 AM

Hi everyone, I'm trying to make a site using Cf pages & workers. I have a SvelteKit app set up correctly and my demo app is running on pages.

Does the adapter automatically take my

*.server.ts

*.server.ts

*.server.ts

*.server.ts and transform it into cf workers?

HHello, I’m Allie!Bindings are generally faster than the API, due to not running over HTTP

B

bun•4/12/24, 12:47 PM

https://tenor.com/view/shocked-surprised-gasp-what-cat-shock-gif-635629308990545194

Tenor

Bbun https://tenor.com/view/shocked-surprised-gasp-what-cat-shock-gif-635629308990545...

H

Hello, I’m Allie!•4/12/24, 12:49 PM

:bibicat:

B

B33pB00p•4/12/24, 1:35 PM

herro all. I have a requirement to geosteer requests to a SIP URI depdending on our cx continent location.

I'm not much of a dev... more like a 50 year old infrastructure script kiddie.

Do you think it's just a matter of altering: https://developers.cloudflare.com/workers/examples/country-code-redirect/

to SIP URIs? Something like,
/
GEO SIP redirector
*
- Redirect users based on their location to different SIP URIs
/

export default {
async fetch(request) {
/**

A map of the SIP URIs to redirect to based on country
@param {Object} countryMap*/const countryMap = {US: "houston2.voip.ms",EU: "london1.voip.ms",CA: "houston2.voip.ms"};// Use the cf object to obtain the country of the requestconst country = request.cf.country;if (country != null && country in countryMap) {const sipUri = countryMap[country];// Remove this logging statement from your final output.console.log(`Based on ${country}-based request, your user would go to ${sipUri}.`);return new Response(null, { status: 302, headers: { "Location": `sip:${sipUri}` }});} else {// Default redirection if country is not foundreturn new Response(null, { status: 302, headers: { "Location": "sip:aaa.voip.ms" }});}},};

?

J

johtso•4/12/24, 2:49 PM

Anyone know of a transactional email sending service with a nice serverless compatible api client?

J

johtso•4/12/24, 2:50 PM

the whole world of api clients is such a mess..

Ggrimpa Can I have multiple consumers in one Worker, each one processing different queue...

J

johtso•4/12/24, 3:05 PM

docs say "one or more" queues https://developers.cloudflare.com/queues/reference/configuration/#workers

J

johtso•4/12/24, 3:11 PM

@João (grimpa) the name of the queue is in the queue property of the message batch https://developers.cloudflare.com/queues/reference/javascript-apis/#messagebatch

Jjohtso Anyone know of a transactional email sending service with a nice serverless comp...

D

DeChill•4/12/24, 3:56 PM

I'm, using https://resend.com/

N

nachotoday•4/12/24, 4:15 PM

Any good resources for learning about system design using the capability security model used by worker RPCs? The Cap'n proto FAQ links out to http://zesty.ca/capmyths/usenix.pdf which I'm working through, but also would like to see some real world examples.

For instance, the RPC announcement gives the example of an auth service that returns a User stub on auth, allowing the caller to interact with the user only if auth succeeds. Let's say we wanted to add a todo list to the app. So, we spin up a few workers:

- web-srv # binds AuthService
  - fetch routes:
    - GET /todos/<todo-list-id> # get the list
    - PUT /todos/<todo-list-id> # set the list
- todo-srv
  - TodoService: Entrypoint # only accessible by auth-srv
    - todoForUser(user_id): TodoUser
  - TodoUser(user_id): RpcTarget 
    - get(id): string[]
    - set(id, string[])
- auth-srv # binds TodoService
  - AuthService: Entrypoint 
    - checkCookie() : User
  - User(user_id): RpcTarget
    - todos(): TodoUser

- web-srv # binds AuthService
  - fetch routes:
    - GET /todos/<todo-list-id> # get the list
    - PUT /todos/<todo-list-id> # set the list
- todo-srv
  - TodoService: Entrypoint # only accessible by auth-srv
    - todoForUser(user_id): TodoUser
  - TodoUser(user_id): RpcTarget 
    - get(id): string[]
    - set(id, string[])
- auth-srv # binds TodoService
  - AuthService: Entrypoint 
    - checkCookie() : User
  - User(user_id): RpcTarget
    - todos(): TodoUser

- web-srv # binds AuthService
  - fetch routes:
    - GET /todos/<todo-list-id> # get the list
    - PUT /todos/<todo-list-id> # set the list
- todo-srv
  - TodoService: Entrypoint # only accessible by auth-srv
    - todoForUser(user_id): TodoUser
  - TodoUser(user_id): RpcTarget 
    - get(id): string[]
    - set(id, string[])
- auth-srv # binds TodoService
  - AuthService: Entrypoint 
    - checkCookie() : User
  - User(user_id): RpcTarget
    - todos(): TodoUser

- web-srv # binds AuthService
  - fetch routes:
    - GET /todos/<todo-list-id> # get the list
    - PUT /todos/<todo-list-id> # set the list
- todo-srv
  - TodoService: Entrypoint # only accessible by auth-srv
    - todoForUser(user_id): TodoUser
  - TodoUser(user_id): RpcTarget 
    - get(id): string[]
    - set(id, string[])
- auth-srv # binds TodoService
  - AuthService: Entrypoint 
    - checkCookie() : User
  - User(user_id): RpcTarget
    - todos(): TodoUser

So a call sequence might be:

- browser: GET /todos/x Cookie=foo
- web-srv: AuthService.checkCookie(foo) -> User
- web-srv: User.todos() -> TodoUser
- web-srv: TodoUser.get(x) # does this go directly to todo-srv or is there a hop through auth-srv?

- browser: GET /todos/x Cookie=foo
- web-srv: AuthService.checkCookie(foo) -> User
- web-srv: User.todos() -> TodoUser
- web-srv: TodoUser.get(x) # does this go directly to todo-srv or is there a hop through auth-srv?

- browser: GET /todos/x Cookie=foo
- web-srv: AuthService.checkCookie(foo) -> User
- web-srv: User.todos() -> TodoUser
- web-srv: TodoUser.get(x) # does this go directly to todo-srv or is there a hop through auth-srv?

- browser: GET /todos/x Cookie=foo
- web-srv: AuthService.checkCookie(foo) -> User
- web-srv: User.todos() -> TodoUser
- web-srv: TodoUser.get(x) # does this go directly to todo-srv or is there a hop through auth-srv?

Maybe I'm overcomplicating this (a simple todo app certainly doesn't need all this), but I'm curious to see what it looks like to go all-in on the capability model.

A

AKStriker98•4/12/24, 4:36 PM

Question since the docs don’t seem to have this clear, is the ja4 fingerprint included in the ja3_hash bot management variable or is this stored in a different variable?

C

cosbgn•4/12/24, 10:16 PM

Concurrent Subrequest Limit

As of this release, we impose a limit on the number of outgoing HTTP requests that a worker can make simultaneously. For each incoming request, a worker can make up to 6 concurrent outgoing fetch() requests.

Concurrent Subrequest Limit

As of this release, we impose a limit on the number of outgoing HTTP requests that a worker can make simultaneously. For each incoming request, a worker can make up to 6 concurrent outgoing fetch() requests.

Concurrent Subrequest Limit

As of this release, we impose a limit on the number of outgoing HTTP requests that a worker can make simultaneously. For each incoming request, a worker can make up to 6 concurrent outgoing fetch() requests.

Concurrent Subrequest Limit

As of this release, we impose a limit on the number of outgoing HTTP requests that a worker can make simultaneously. For each incoming request, a worker can make up to 6 concurrent outgoing fetch() requests.

This means that something like this:

await Promise.all(jobs.map(async (job_url) => await $fetch(job_url))

await Promise.all(jobs.map(async (job_url) => await $fetch(job_url))

await Promise.all(jobs.map(async (job_url) => await $fetch(job_url))

await Promise.all(jobs.map(async (job_url) => await $fetch(job_url))

Would take very long right?
It would do 6 + 6 + 6 + 6 until it reaches the limit of 50 per request

J

Jonah•4/12/24, 10:20 PM

Does

*domain.com/*

*domain.com/*

*domain.com/*

*domain.com/*

in a Cloudflare worker trigger, no longer process all requests? A few months ago, that trigger pattern ran on every request, domain.com/style.css and so on, now it only runs on / ?

Ccosbgn ``` Concurrent Subrequest Limit As of this release, we impose a limit on the nu...

K

kian•4/12/24, 10:27 PM

Not quite 6+6+6+6

K

kian•4/12/24, 10:27 PM

Once one finishes, another one can start

K

kian•4/12/24, 10:27 PM

It's just 6 concurrent, not batches of 6

Ccosbgn ``` Concurrent Subrequest Limit As of this release, we impose a limit on the nu...

B

bun•4/13/24, 12:32 AM

https://flareutils.pages.dev/promisequeue/ there's an util developed for this specific usecase. runs 6 promises concurrently

PromiseQueue | Flareutils

Small Utilities and little goodies that make developing with Cloudflare easier and faster.

D

dom•4/13/24, 7:58 AM

Workers wrangler logs and realtime logs are not returning anything for a specified worker. Is anyone else experiencing these issues?

D

DeChill•4/13/24, 4:19 PM

Hi, is there any way to trigger workers in response to Messages from an external Kafka topic?

O

oof2win2•4/13/24, 4:58 PM

i dont know if im going insane at this point, but i have this weird scenario. i have a public api and a discord bot that both have the same functionality essentially, so i was thinking that i would take out the core api functionality, put it into a separate worker and make the api and discord bot have bindings to the core functions

O

oof2win2•4/13/24, 4:59 PM

does this make sense? like the main benefit would be that the discord bot wouldnt be using http calls to communicate with the worker

Ooof2win2 does this make sense? like the main benefit would be that the discord bot wouldn...

B

Beny•4/13/24, 5:00 PM

Yes that would be good, no point having 2 copies of the same code

O

oof2win2•4/13/24, 5:01 PM

hmm ill think it over

Question: if a worker exceeds its 6 simultaneous network connection count, is it the case that it wi

Similar Threads

Similar Threads

Similar Threads