How can I get my worker to run more than 30 seconds? I'm having a cron job trigger a database syncin

How can I get my worker to run more than 30 seconds?
I'm having a cron job trigger a database syncing function, which can run for 30mins from the start

Devon•3/14/24, 1:19 PM

can you poll the DB sync status periodically? serverless isn't really meant for long running jobs

DDevon can you poll the DB sync status periodically? serverless isn't really meant for...

JoeDoeOP•3/14/24, 1:20 PM

yeah it's possible I guess...

JoeDoeOP•3/14/24, 1:20 PM

and would need to trigger it more frequently then

JoeDoeOP•3/14/24, 1:20 PM

and sync smaller bits at a time

Devon•3/14/24, 1:20 PM

serverless isn't meant for running 30 minute tasks

JJoeDoe What will happen if a cron is triggered, and it's a long running process, and th...

JoeDoeOP•3/14/24, 1:21 PM

do you know about this?

Devon•3/14/24, 1:21 PM

so I'd see if you can run it asyncronously. or look at stateful compute options

Devon•3/14/24, 1:21 PM

oh you can setup a job state table/KV to check job status before runnign thing. pretty common pattern overall.

JJoeDoe How can I get my worker to run more than 30 seconds? I'm having a cron job trigg...

Chaika•3/14/24, 1:27 PM

sounds like you're hitting this:

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.

Chaika•3/14/24, 1:27 PM

although that mentions Unbound, I would imagine Standard would be the same

AnotherAlex•3/14/24, 1:55 PM

Is there any way to check the compatibility version/flags on a deployed worker? I can't use the Header.getSetCookie function because it doesn't exist, although the code editor reports it existing on the type.

kian•3/14/24, 1:58 PM

You could hit the API to see what it's currently set to, but that's about it.

kian•3/14/24, 1:59 PM

The types aren't aware of your compat flag unless you set them up to point to a specific date, so it probably is always showing you the latest

Hunter Bertoson•3/14/24, 2:12 PM

If I am using the mailchannels API in my worker to send emails. Does my domain need to be on Cloudflare? Or does it just need have the DNS Records?

Mitya•3/14/24, 2:23 PM

Is something up with Workers today? I test locally, but with the remoteremote flag, and in my console output I'm getting lots of [object Object][object Object] references for things that are definitely not, 100% note, objects - they're just strings. This is shown by instead logging the result of encodeURIComponent(myStr)encodeURIComponent(myStr), and the problem goes away. Obviously it's hard to give a reproducible example, but is there anything that might be causing this? It's just started happening today, in multiple endpoints of my code, and wasn't happening before.

Mitya•3/14/24, 2:24 PM

The objects are illusory; when I ask my PG util what query it's about to run, it says it contains [object Object][object Object]. This would obviously result in a query failure, but the query runs fine, which again suggests that this is some sort of illusion in Workers or the Workers console output.

BlackPearl•3/14/24, 2:40 PM

anyone know how to use mongodb in normal js for cloudflare worker

Kkian You could hit the API to see what it's currently set to, but that's about it.

AnotherAlex•3/14/24, 3:14 PM

Thanks. They are both empty but I think a downstream service might be affecting compatibility so going to try updating that.

Enduriel•3/14/24, 3:24 PM

Is there a way to measure the ~CPU time of a request when developing locally?

Enduriel•3/14/24, 3:24 PM

I'd like to know the approximate ballpark of w/e I'm doing

BBlackPearl anyone know how to use mongodb in normal js for cloudflare worker

Isaac McFadyen•3/14/24, 4:31 PM

You cannot currently connect to MongoDB from a Cloudflare Worker except through the MongoDB Atlas Data API.

Isaac McFadyen•3/14/24, 4:32 PM

You might also be able to use Prisma Accelerate but I haven't tried it.

JoeDoeOP•3/14/24, 5:41 PM

so the D1 database can only be accessed via workers?

JJoeDoe so the D1 database can only be accessed via workers?

Isaac McFadyen•3/14/24, 5:43 PM

D1 can be accessed either from Workers or using the Cloudflare API (which has rate limits so the Workers binding is better)

IIsaac McFadyen D1 can be accessed either from Workers or using the Cloudflare API (which has ra...

JoeDoeOP•3/14/24, 5:44 PM

roger

JoeDoeOP•3/14/24, 6:55 PM

ok, so lets say I use cron jobs to slowly sync my database, and then I want to run some script to process the data in the D1 database, what would be the best way to do that?

JJoeDoe ok, so lets say I use cron jobs to slowly sync my database, and then I want to r...

Devon•3/14/24, 7:20 PM

IMHO that is way too vague for anyone to provide useful direction

Hunter Bertoson•3/14/24, 7:51 PM

I have a worker set up to send emails using the MailChannels Integration. It works great from the console. But when I call it from my localhost to try and send an email. I get a 403 error. Is this not possible? My site is Astro and is deployed on Cloudflare Pages.

Devon•3/14/24, 7:54 PM

when you say localhost are you running the Worker locally? if so that won't work, as the MC integration only allows calls from deployed Workers IPs

Hunter Bertoson•3/14/24, 7:55 PM

I am not. I can try that. I did try it on my deployed site and it still came back with a 403 error though

Devon•3/14/24, 7:57 PM

okay yeah it has to be running deployed in CF env. I just went through getting MC working (using the free setup with Workers). It's sort of a pita;) You did the Domain Lockdown DNS thing, DKIM, SPF, etc...?

Devon•3/14/24, 7:57 PM

and they usually send an error messsage back in the response body that has more info

Hunter Bertoson•3/14/24, 7:58 PM

Yeah. I have the Domain lockdown DNS, DKIM, SPF all that Jazz. If I run it from the Dashboard. It works great. Let me try it again with some better errror handling

Devon•3/14/24, 7:59 PM

I guess I'm not sure I understand Dashboard vs "localhost" for the execution if the Worker is running deployed (not locally).

Hunter Bertoson•3/14/24, 7:59 PM

Here:

Devon•3/14/24, 8:01 PM

ah! okay. but when you trigger the Worker by calling it's endpoint it fails? that IS odd

Hunter Bertoson•3/14/24, 8:09 PM

I figured it out now.

Devon•3/14/24, 8:10 PM

what was it?

Hunter Bertoson•3/14/24, 8:11 PM

Hunter Bertoson•3/14/24, 8:13 PM

I think I want the "cf-worker" Header?

Devon•3/14/24, 8:13 PM

is it being called in a way that would pass the referrer? CORS works off the Origin header (set by browsers when executing AJAX calls). So maybe you just need a CORS implementation?

Hunter Bertoson•3/14/24, 8:14 PM

Not sure. i will take a look when I get back. Got to go some Ref some soccer games. Thanks for the help!

kian•3/14/24, 10:13 PM

Sounds like the dashboard is wrong

kian•3/14/24, 10:27 PM

Services is https://developers.cloudflare.com/workers/configuration/bindings/about-service-bindings/, not Tail Workers.

JoeDoeOP•3/15/24, 6:50 AM

Is there any sort of "on exit" or "on close" event callback for workers?

How can I get my worker to run more than 30 seconds? I'm having a cron job trigger a database syncin

Similar Threads

Similar Threads

Similar Threads