Is there way to rerun cron job if it fails overriding the scheduled trigger.

Ff8 Is there way to rerun cron job if it fails overriding the scheduled trigger.

C

Chaika•6/12/24, 7:46 PM

you can manually call the scheduled handler from a fetch event, ex:

   if (newurl.pathname === "/supersecretfetchmetotriggercron" && request.headers.has("APIKEY") && request.headers.get("APIKEY") === env.CRONAPIKEY) {
// if you're using any properties of the event like cron/etc you can recreate it instead of passing null
      await this.scheduled(null, env, ctx);
      return new Response("ok - triggered");
    }

   if (newurl.pathname === "/supersecretfetchmetotriggercron" && request.headers.has("APIKEY") && request.headers.get("APIKEY") === env.CRONAPIKEY) {
// if you're using any properties of the event like cron/etc you can recreate it instead of passing null
      await this.scheduled(null, env, ctx);
      return new Response("ok - triggered");
    }

   if (newurl.pathname === "/supersecretfetchmetotriggercron" && request.headers.has("APIKEY") && request.headers.get("APIKEY") === env.CRONAPIKEY) {
// if you're using any properties of the event like cron/etc you can recreate it instead of passing null
      await this.scheduled(null, env, ctx);
      return new Response("ok - triggered");
    }

   if (newurl.pathname === "/supersecretfetchmetotriggercron" && request.headers.has("APIKEY") && request.headers.get("APIKEY") === env.CRONAPIKEY) {
// if you're using any properties of the event like cron/etc you can recreate it instead of passing null
      await this.scheduled(null, env, ctx);
      return new Response("ok - triggered");
    }

CChaika you can manually call the scheduled handler from a fetch event, ex: ```js if ...

F

f8OP•6/12/24, 7:47 PM

Gotcha, how would i know it fails to then call this

Ff8 Gotcha, how would i know it fails to then call this

C

Chaika•6/12/24, 7:47 PM

oh you mean automatically retry on failure?

CChaika oh you mean automatically retry on failure?

F

f8OP•6/12/24, 7:47 PM

yes

F

f8OP•6/12/24, 7:48 PM

my cron runs every hour. if it fail i want it to run right after again

C

Chaika•6/12/24, 7:48 PM

nothing native, I mean you could just extract your cron logic to a different function and try.. cronfunc() catch.. cronfunc() but a bit silly lol, it usually makes more sense to specifically retry what could fail/scope your retries better

C

Chaika•6/12/24, 7:48 PM

does your cron rely on a third party api fetch that can fail? Just have that part retry, etc

C

Chaika•6/12/24, 7:50 PM

of course there are specific conditions for the retry being able to work like that. If you run out of resources (cpu, mem), it's just going to fail

J

johtso•6/12/24, 9:51 PM

is there anything that will show what your workers usage is as a proportion of your plan's included usage?

J

johtso•6/12/24, 9:51 PM

those two things seem rather disjointed at the moment

J

johtso•6/12/24, 9:52 PM

even a 3rd party tool

J

johtso•6/12/24, 9:52 PM

I know I'm way bellow the allowance, but it still freaks me out not being able to see it

S

samuelarogbonlo•6/12/24, 10:09 PM

Hello Folks,

I'm having an issue with my Cloudflare workers and need some help. I have two workers in different Cloudflare accounts:

Worker 1 in Account 1 with the domain `example.com` and routes:
- test.example.com/list/fruits
- test.example.com/list/animals
  - Worker 2 in Account 2 with the domain `myexample.com` and routes:
  - test.myexample.com/list/fruits
  - test.myexample.com/list/animals
    I want to update Worker 2 to use the routes from Worker 1. How can I configure cross-account worker domain access to achieve this?

J

johtso•6/12/24, 10:58 PM

maybe you could proxy requests to the other domain using a worker?

F

f8OP•6/13/24, 12:28 AM

when i run my cronjob locally via curl "http://localhost:56968/__scheduled?cron=0+*+*+*+*" it begins running what's in my async scheduled { } body. After about a minute it reruns eventhough im setting cron to 0 * * * *

F

f8OP•6/13/24, 12:28 AM

Why is this the case?

F

f8OP•6/13/24, 4:04 AM

Are cron's cpu limit 15min or 30s for standard plan when set to run exactly once every hour? Running locally mine I get a cpu time limit after 50000ms but wondering if this is just because it's not given 15min locally because it can be ran more than every hour?

F

f8OP•6/13/24, 4:23 AM

"On the Unbound billing model, scheduled Workers (Cron Triggers) have different limits on CPU time based on the schedule interval. When the schedule interval is less than 1 hour, a Scheduled Worker may run for up to 30 seconds. When the schedule interval is more than 1 hour, a scheduled Worker may run for up to 15 minutes."

Isn't unbound model deprecated and now is standard model? Does this 15 minutes still apply? Also what happens if it runs exactly every hour. Technically says "more than 1 hour" but what if it's every one hour

N

nehatkhan•6/13/24, 4:23 AM

I have created a Hono js app with node js adapter, how to deploy the app to cloudfare workers. Please help

Nnehatkhan I have created a Hono js app with node js adapter, how to deploy the app to clou...

S

Sam•6/13/24, 8:39 AM

https://developers.cloudflare.com/workers/wrangler/commands/#deploy

Cloudflare Docs

Commands - Wrangler · Cloudflare Workers docs

Create, develop, and deploy your Cloudflare Workers with Wrangler commands.

C

cosbgn•6/13/24, 10:47 AM

What's canceled?

E

Erisa•6/13/24, 11:03 AM

The user/client making the request cancelled the request before it could complete

E

Erisa•6/13/24, 11:04 AM

you would get this if you for example ran

curl

curl

curl

curl against a url and then did ctrl+c to cancel it before it finished, or loaded a webpage and then closed the tab or pressed the X to stop loading before it finished

C

cosbgn•6/13/24, 11:05 AM

Ok thanks

C

cosbgn•6/13/24, 11:18 AM

@Erisa I'm sending a cron, and cf is returning a 524 and "canceled" - Is this a timeout from workers or from my cron provider?

C

cosbgn•6/13/24, 11:18 AM

the cron job is not from cf, it's another platform

C

cosbgn•6/13/24, 11:19 AM

Seems to die at 100 seconds exact

C

cosbgn•6/13/24, 11:27 AM

even if the cron calls directly /api ? does it still pass through the cdn?

C

cosbgn•6/13/24, 11:28 AM

so workers practically can't run longer than 100 seconds unless send back a stream

C

cosbgn•6/13/24, 11:28 AM

or maybe if I move the logic inside waitUntill it could run without the 100 second limit?

C

cosbgn•6/13/24, 11:29 AM

or maybe instead of relying on http crons I could use cloudflare ones, that should avoid the 100 second response issue

C

cosbgn•6/13/24, 11:30 AM

So the best solution is using cloudflare cron triggers, right?

C

cosbgn•6/13/24, 11:31 AM

Ok I'll check those out. Thanks!

J

jambaloq•6/13/24, 1:23 PM

Hello,
is this possible to add a workers route to a different zone on different CF account?
I see just those domains (zones) that exist in the same CF account

J

jambaloq•6/13/24, 1:24 PM

got it, thanks for the prompt response!

M

Mr. ab1t•6/13/24, 6:46 PM

Hello awesome people. I'm trying to write simple worker, which once called it's API it get a parameter and based on that call some another webpage (outside to cloudflare - weather sites), parse that response and sent to the original caller. But ... still getting internal error. Are there some restrictions what workers can/can't ie. firing POSTs to different sites?

MMr. ab1t Hello awesome people. I'm trying to write simple worker, which once called it's ...

I

Isaac McFadyen•6/13/24, 11:30 PM

You'll want to check your Worker logs to see if there's an error listed, which you can do either with the

wrangler tail

wrangler tail

wrangler tail

wrangler tail command or on the Cloudflare dashboard.

I

Isaac McFadyen•6/13/24, 11:30 PM

It's possible there's an error in your code, or perhaps the external weather API uses a non-standard port which Workers doesn't support.

S

similardisaster•6/14/24, 7:37 PM

is the best way to conditionally let a request just "pass through" to orign by using "fetch(request)"? Will that just pass the request through as if the worker was not there?

E

Erisa•6/14/24, 7:39 PM

return fetch(request)

return fetch(request)

return fetch(request)

return fetch(request) should do that, yeah

E

Erisa•6/14/24, 7:39 PM

I also think if you dont return anything then it does it automatically

but I havent confirmed that recently with module syntax and such

S

similardisaster•6/14/24, 7:40 PM

thanks! interesting

EErisa I also think if you dont return anything then it does it automatically 🤔 but I ...

E

Erisa•6/14/24, 7:40 PM

Ah seems this doesnt happen in module syntax, you instead get

Uncaught (in response) TypeError: Incorrect type for Promise: the Promise did not resolve to 'Response'.

Uncaught (in response) TypeError: Incorrect type for Promise: the Promise did not resolve to 'Response'.

Uncaught (in response) TypeError: Incorrect type for Promise: the Promise did not resolve to 'Response'.

Uncaught (in response) TypeError: Incorrect type for Promise: the Promise did not resolve to 'Response'.

E

Erisa•6/14/24, 7:41 PM

So yeah go with

return fetch(request)

return fetch(request)

return fetch(request)

return fetch(request), it should be identical to not having the worker

EErisa Ah seems this doesnt happen in module syntax, you instead get `Uncaught (in resp...

E

Erisa•6/14/24, 7:42 PM

Though ironically since this is an exception it does pass through if you execute

ctx.passThroughOnException()

ctx.passThroughOnException()

ctx.passThroughOnException()

ctx.passThroughOnException() before returning nothing, but I dont recommend doing that

E

Erisa•6/14/24, 7:43 PM

ctx.passThroughOnException()

ctx.passThroughOnException()

ctx.passThroughOnException()

ctx.passThroughOnException() is pretty useful if you dont want errors in your worker to affect traffic - I only dont recommend it if you plan to return nothing lol

S

similardisaster•6/14/24, 7:43 PM

am using it, was just curious if they provided a similar "passThrough()" method or equivelant

E

Erisa•6/14/24, 7:44 PM

Makes sense, yeah I've always seen and done

return fetch(request)

return fetch(request)

return fetch(request)

return fetch(request) for that

EErisa `ctx.passThroughOnException()` is pretty useful if you dont want errors in your ...

C

Chaika•6/14/24, 9:15 PM

may be worth mentioning I found out the hard way that's not foolproof lol, if you consume the body (of a request that does have a body) before the worker dies you're dead in the water

C

Chaika•6/14/24, 9:16 PM

oh wow 3 weeks ago they added that warning https://github.com/cloudflare/cloudflare-docs/commit/8fc09ccf816f7417d3e3d46852f74f34834ea15b
I just got unlucky when I tried it a while ago and found that out the hard way lol

Is there way to rerun cron job if it fails overriding the scheduled trigger.

Similar Threads

Similar Threads

Similar Threads