with a worker in the middle, you gotta handle frontend <-> worker <-> supa errors

Z

zachOP•3/29/24, 4:37 AM

i think if you have properly scoped access tokens for supa, you'll be chill to use it on the clientside

Z

zachOP•3/29/24, 4:38 AM

that said, I've never been happy with letting untrusted internet interface directly with my db

A

Alvee•3/29/24, 4:43 AM

does cloudflare loadbalacner have concurrent connectino limits ?

V

vimo•3/29/24, 11:23 AM

Is there a way to pause the quick reload of the workers playground? or add a delay to its reload?

K

K•3/29/24, 11:54 AM

There is an issues with concurrent fetches and queing. Its described and posted with workaround here workers-and-pages-helpSimultaneous fetches over 6 are being dropped instead of being queued.

I don't know whom to ping for this (which channel) so I'll leave it here.

Zzach with a worker in the middle, you gotta handle frontend <-> worker <-> supa error...

F

Foreignity•3/29/24, 12:41 PM

Thank you @neu² for your answers. Do you know if the supabase integrations to workers include the auth?

MMurder Chicken Thanks again kian. I submitted a request and will begin to look into Tail Worke...

M

marina.lipeck•3/29/24, 1:15 PM

Hey! were you able to get an increase through form?

B

bkyerv•3/29/24, 2:08 PM

hello everyone. has anyone successfully used googleapis lib with cloudflare to interact with products like google sheets or similar?

FForeignity Thank you @neu² for your answers. Do you know if the supabase integrations to wo...

Z

zachOP•3/29/24, 2:13 PM

You'd do it through their js sdk, so yes

Bbkyerv hello everyone. has anyone successfully used googleapis lib with cloudflare to i...

C

Chaika•3/29/24, 2:17 PM

Workers aren't Nodejs, they have some node compat but not full, iirc googleapis uses a ton of node apis including their own version of axios, would have to change it pretty majorly

N

Nehal Ali•3/29/24, 2:38 PM

Hello , i have a question regarding caching through fetch API : https://developers.cloudflare.com/workers/examples/cache-using-fetch/
let's say i have a worker that i want to fetch data from and cache this data for an hour so that the next request the data should be taken from the cache.
So i am expecting for the same request with in my window of an hr, the 2nd fetch response would be taken from the cache not go to make another http request correct ?
if this is the case why for every fetch request i do the worker gets called and returns the response?
How can i validate that this is from the cache or not?
i have set

cf: {
        cacheTtl: 3600,
        cacheEverything: true,
      },

cf: {
        cacheTtl: 3600,
        cacheEverything: true,
      },

cf: {
        cacheTtl: 3600,
        cacheEverything: true,
      },

cf: {
        cacheTtl: 3600,
        cacheEverything: true,
      },

in my fetch request.
is there somethig i should add ?

Cloudflare Docs

Cache using fetch · Cloudflare Workers docs

Setting the cache level to Cache Everything will override the default cacheability of the asset. For time-to-live (TTL), Cloudflare will still rely on …

NNehal Ali Hello , i have a question regarding caching through fetch API : https://develope...

C

Chaika•3/29/24, 2:41 PM

So i am expecting for the same request with in my window of an hr, the 2nd fetch response would be taken from the cache not go to make another http request correct ?

Cache by default is per colo/cf dc, and some Cloudflare Points of Presence even have multiple colos, so you might be routed to a different one/different visitors will certainly be, and get MISS's, plus TTL is just a max and CF may evict it far before then based on access frequency.

if this is the case why for every fetch request i do the worker gets called and returns the response?

The worker will always be invocation/run on every request, even if using cache

How can i validate that this is from the cache or not?

Check response headers for CF-Cache-Status

CChaika Workers aren't Nodejs, they have some node compat but not full, iirc googleapis ...

I

Isaac McFadyen•3/29/24, 3:39 PM

IIRC they also use gRPC which isn't supported even with polyfills

I

Isaac McFadyen•3/29/24, 3:39 PM

Or at least a lot of them do

I

Isaac McFadyen•3/29/24, 3:39 PM

Or maybe that's just GCP's?

R

Razzmatazz•3/29/24, 5:00 PM

Anyone know why purging by URL via a fetch request to the API (https://api.cloudflare.com/client/v4/zones/[zone]/purge_cache) done via a worker apparently is ignored? The API response indicates success, but the cache is definitely not purged.

R

rootalien•3/29/24, 6:35 PM

R

rootalien•3/29/24, 6:36 PM

When I open it from the incognito tab, cf-ray id is removed from AMS and the problem is solved.

R

rootalien•3/29/24, 6:36 PM

@Chaika

Rrootalien @Chaika

C

Chaika•3/29/24, 6:39 PM

why tag me?

R

rootalien•3/29/24, 6:39 PM

I thought it was an authority, sorry.

R

rootalien•3/29/24, 6:45 PM

I don't know but I use it for a fee.

Original message was deleted

S

sirhype•3/29/24, 10:49 PM

There shouldn't be any, you can find the list of limits here -> https://developers.cloudflare.com/workers/platform/pricing/

H

Hanging_on_edge•3/30/24, 10:04 AM

I am trying to use discord.js in my bot to create a feature which can kick off a user
I wanted to know if it is compatible with cloudfare workers ?

H

Hanging_on_edge•3/30/24, 10:05 AM

I want to use the npm package discord.js

Original message was deleted

H

Hanging_on_edge•3/30/24, 10:10 AM

can you please elaborate

H

Hanging_on_edge•3/30/24, 10:10 AM

\

R

Razzmatazz•3/30/24, 10:23 AM

A longer explanation is that cloudflare workers are meant to run bits of code and then terminate. A discord bot needs to continue running indefinitely.

RRazzmatazz A longer explanation is that cloudflare workers are meant to run bits of code an...

H

Hello, I’m Allie!•3/30/24, 10:36 AM

Well, that’s not entirely the case. It is more that Discord has blocked Worker’s IPs from connecting to the Gateway

H

Hello, I’m Allie!•3/30/24, 10:37 AM

You could build a Gateway Bot with Durable Objects. It wouldn’t be quite as ergonomic as one built in Node, but it definitely is doable

H

Hello, I’m Allie!•3/30/24, 10:37 AM

Other than that it wouldn’t actually connect

HHello, I’m Allie!Well, that’s not entirely the case. It is more that Discord has blocked Worker’s...

R

Razzmatazz•3/30/24, 11:23 AM

Good point. Multiple reasons it’s not possible to use the discord.js package to build a bit in a worker.

RRazzmatazz Good point. Multiple reasons it’s not possible to use the discord.js package to ...

H

Hanging_on_edge•3/30/24, 12:23 PM

can you please give other reasons too

HHanging_on_edge can you please give other reasons too

R

Razzmatazz•3/30/24, 12:27 PM

How many independently-disqualifying reasons do you need for why it's not feasible to use discord.js to make a discord bot in a worker?

V

vloe•3/30/24, 1:28 PM

i'm so lost...

V

vloe•3/30/24, 1:29 PM

added cors to my cf-worker but everybody can still access my endpoint:

V

vloe•3/30/24, 1:30 PM

#[event(fetch)]
async fn fetch(req: HttpRequest, _env: Env, _ctx: Context) -> Result<Response<Body>> {
    console_error_panic_hook::set_once();

    let cors = CorsLayer::new()
        .allow_origin("https://shorter.dev".parse::<HeaderValue>().unwrap())
        .allow_headers([CONTENT_TYPE])
        .allow_methods([Method::GET, Method::POST, Method::HEAD, Method::OPTIONS]);

    let app = Router::new()
        .route("/", get(|| async { "shorter.dev server!" }))
        .merge(mount())
        .layer(cors)
        .call(req)
        .await?;

    Ok(app)
}

#[event(fetch)]
async fn fetch(req: HttpRequest, _env: Env, _ctx: Context) -> Result<Response<Body>> {
    console_error_panic_hook::set_once();

    let cors = CorsLayer::new()
        .allow_origin("https://shorter.dev".parse::<HeaderValue>().unwrap())
        .allow_headers([CONTENT_TYPE])
        .allow_methods([Method::GET, Method::POST, Method::HEAD, Method::OPTIONS]);

    let app = Router::new()
        .route("/", get(|| async { "shorter.dev server!" }))
        .merge(mount())
        .layer(cors)
        .call(req)
        .await?;

    Ok(app)
}

#[event(fetch)]
async fn fetch(req: HttpRequest, _env: Env, _ctx: Context) -> Result<Response<Body>> {
    console_error_panic_hook::set_once();

    let cors = CorsLayer::new()
        .allow_origin("https://shorter.dev".parse::<HeaderValue>().unwrap())
        .allow_headers([CONTENT_TYPE])
        .allow_methods([Method::GET, Method::POST, Method::HEAD, Method::OPTIONS]);

    let app = Router::new()
        .route("/", get(|| async { "shorter.dev server!" }))
        .merge(mount())
        .layer(cors)
        .call(req)
        .await?;

    Ok(app)
}

#[event(fetch)]
async fn fetch(req: HttpRequest, _env: Env, _ctx: Context) -> Result<Response<Body>> {
    console_error_panic_hook::set_once();

    let cors = CorsLayer::new()
        .allow_origin("https://shorter.dev".parse::<HeaderValue>().unwrap())
        .allow_headers([CONTENT_TYPE])
        .allow_methods([Method::GET, Method::POST, Method::HEAD, Method::OPTIONS]);

    let app = Router::new()
        .route("/", get(|| async { "shorter.dev server!" }))
        .merge(mount())
        .layer(cors)
        .call(req)
        .await?;

    Ok(app)
}

Vvloe ```rust #[event(fetch)] async fn fetch(req: HttpRequest, _env: Env, _ctx: Contex...

H

Hello, I’m Allie!•3/30/24, 1:32 PM

CORS does not prevent those that don't respect CORS from hitting your Worker, if that is what you mean

H

Hello, I’m Allie!•3/30/24, 1:32 PM

It only works for stuff like Browsers

V

vloe•3/30/24, 1:38 PM

thank you. any idea how i can prevent everyone from hitting my worker except my own requests?

Vvloe thank you. any idea how i can prevent everyone from hitting my worker except my ...

H

Hello, I’m Allie!•3/30/24, 1:39 PM

You can check certain headers to see if they match, but if the attackers are persistent, there isn't much you can do

H

Hello, I’m Allie!•3/30/24, 1:40 PM

There isn't a 100% foolproof way to ensure that a request is only coming from your app(unless you control the client, and don't allow anyone access to it)

V

vloe•3/30/24, 1:46 PM

i see, thank you!

A

Ashley•3/30/24, 1:53 PM

Is there a knack to getting service bindings working locally? I have two running locally, both using

npm run dev

npm run dev

and worker A is bound to worker B. Worker B calls Worker A, and when I start Worker B I get:

⛅️ wrangler 3.41.0
-------------------
▲ [WARNING] This worker is bound to live services: AUTH (authentication-worker)


Your worker has access to the following bindings:
- Services:
  - AUTH: authentication-worker
▲ [WARNING] ⎔ Support for service bindings in local mode is experimental and may change.

⛅️ wrangler 3.41.0
-------------------
▲ [WARNING] This worker is bound to live services: AUTH (authentication-worker)


Your worker has access to the following bindings:
- Services:
  - AUTH: authentication-worker
▲ [WARNING] ⎔ Support for service bindings in local mode is experimental and may change.

⛅️ wrangler 3.41.0
-------------------
▲ [WARNING] This worker is bound to live services: AUTH (authentication-worker)


Your worker has access to the following bindings:
- Services:
  - AUTH: authentication-worker
▲ [WARNING] ⎔ Support for service bindings in local mode is experimental and may change.

⛅️ wrangler 3.41.0
-------------------
▲ [WARNING] This worker is bound to live services: AUTH (authentication-worker)


Your worker has access to the following bindings:
- Services:
  - AUTH: authentication-worker
▲ [WARNING] ⎔ Support for service bindings in local mode is experimental and may change.

The warning itself is a little confusing, as it's running locally, so I don't know why it's mentioning live. Either way, when I hit it with a request, I get

Cannot read properties of undefined (reading 'fetch')

Cannot read properties of undefined (reading 'fetch')

Cannot read properties of undefined (reading 'fetch')

Cannot read properties of undefined (reading 'fetch')

if I run Worker B in remote mode, it works fine (but presumably hits my live worker), so I'm quite confused. Any ideas?

A

Ashley•3/30/24, 1:53 PM

(both are using the same version of Wrangler)

RRazzmatazz How many independently-disqualifying reasons do you need for why it's not feasib...

H

Hanging_on_edge•3/30/24, 2:06 PM

Sorry I couldnt understand thats why

AAshley Is there a knack to getting service bindings working locally? I have two running...

A

Ashley•3/30/24, 2:10 PM

Okay I just read https://github.com/cloudflare/workers-sdk/issues/1182, in which case https://developers.cloudflare.com/workers/configuration/bindings/about-service-bindings/#local-development is highly misleading

C

Cretezy•3/30/24, 3:32 PM

Is it possible to see the cron logs when an error happens?

CCretezy Is it possible to see the cron logs when an error happens?

H

Hello, I’m Allie!•3/30/24, 3:33 PM

Only if you have a tail session/logpush job running

C

Cretezy•3/30/24, 3:44 PM

Alright thanks. I think capturing the error message when an exception is thrown inside a cron worker and displaying it inside the log cron triggers would be very useful.

with a worker in the middle, you gotta handle frontend <-> worker <-> supa errors

Similar Threads

Similar Threads

Similar Threads