That page covers the Node compatibility layer. Some modules are supported, but Node as a whole isn’t
That page covers the Node compatibility layer. Some modules are supported, but Node as a whole isn’t
0Thanks for clarifying
I was told by multiple people that I could run langchain with Cloudflare workers & wrangler
Am I being led in the wrong direction?
The issue I’m having is that I want to run langchain in a react native mobile application but the Hermes engine does not use “readable streams” which is how langchain responds
I was told by multiple people that I could run langchain with Cloudflare workers & wrangler
Am I being led in the wrong direction?
The issue I’m having is that I want to run langchain in a react native mobile application but the Hermes engine does not use “readable streams” which is how langchain responds
HYou can run some models on Cloudflare with https://ai.cloudflare.com, but as to whether you can run Langchain itself, I'm not sure
MGuys, adding Durable Objects to my code causes it to 200ms-1s additional response times, Is that expected? Was trying to replace KV for authentication management
MEven when setting the locationHint
Swhat country
MBrazil, was trying to set locationHint to sam (South America)
CI'm pretty sure there is no Durable Object in South America 
This cool page shows where DOs are created: https://where.durableobjects.live
This cool page shows where DOs are created: https://where.durableobjects.live
Tracking where Durable Objects are created, wherever you are in the world.
MThat makes sense now! Thank you guys
MWill stick with KV for now, and durable for ws connections
H
GHi! The cloudflare API permit to create a Pages deployment from a repository, but I don't see any way from the API to do the same from a « Direct Upload » (.zip file) https://developers.cloudflare.com/api/operations/pages-deployment-create-deployment
Is it planned?
Is it planned?
Interact with Cloudflare's products and services via the Cloudflare API
CCan I fetch another worker in the same zone if my worker runs on a custom domain?
CLast I heard, you should be able to, the docs say so, but you can't unless you remove the miminimal_subrequests flag, ex: https://github.com/cloudflare/workerd/issues/787#issuecomment-1597396851
Service Bindings may be a better fit unless you really need the same-zone fetch: https://developers.cloudflare.com/workers/configuration/bindings/about-service-bindings/
They don't go over the internet, same isolate, so generally faster, and you can pass special properties over the request.cf property
Service Bindings may be a better fit unless you really need the same-zone fetch: https://developers.cloudflare.com/workers/configuration/bindings/about-service-bindings/
They don't go over the internet, same isolate, so generally faster, and you can pass special properties over the request.cf property
Service bindings are an API that facilitate Worker-to-Worker communication via explicit bindings defined in your configuration.
GitHub
Which Cloudflare product(s) does this pertain to? Workers/Other What version of Wrangler are you using? 2.9.0 What operating system are you using? Mac Describe the Bug Same zone worker <-> wo...
CThanks, that is what I am looking for. I encountered the same 522 error issue.
BI have 400 users details like email, name, contact etc inside an array. I want to insert/update all these 400 users profiles information inside 3rd party service called klaviyo for email marketing. In adding 1 user information into this 3rd party API, it takes 2 fetch requests (max 2 for update). So for 400 users its taking 800 fetch requests at a time for inserting. I am getting error inside logs that "subrequest limit exceeds" means so many subrequest a time...
What is solution of this, can anyone please help?
NOTE: I am using premium version not free
I have also attached a screenshot of my code with result array of 400 users and then adding them one by one into klaviyo
What is solution of this, can anyone please help?
NOTE: I am using premium version not free
I have also attached a screenshot of my code with result array of 400 users and then adding them one by one into klaviyo
AEither limit the number of customers you send per scheduled run (e.g. in your DB, have a column that is "sent" or whatever) and then retrieve X number per scheduler run and only pull ones where they aren't sent yet to the API. Or use https://developers.cloudflare.com/queues/, and push each customer as a message into a queue, and have a separate worker process the queue
AWell, those are two options anyway
AAI also don't know what it is for the new pricing model, as they only talk about bound vs unbound there
AAs there is a big difference (50 for bound, 1000 for unbound) but not sure what it will be for standard (the new pricing model that is now available)
UIf you change to the new billing model (Standard) you can do 1,000 fetch requests. You could also send all of them to queues instead so you can support retrying the requests over time if needed.
AThat answers my Q on the standard model, thanks 
BI tried a way called batch processing.... I first add 20 users then added delay function (Promise) wait of 60 seconds or setInterval/setTimeout and then add another 20 users...
but it didn't worked, dont know why? It always adding that first 20 users only after every 60 seconds...not going to next 20 users.
Added 2 screenshots of delay promise function and setTimeout, but both not worked. always adding/updating same chunk/batch. Whats reason?
but it didn't worked, dont know why? It always adding that first 20 users only after every 60 seconds...not going to next 20 users.
Added 2 screenshots of delay promise function and setTimeout, but both not worked. always adding/updating same chunk/batch. Whats reason?
BWill try your above approaches also, that queue and limits thanks...but above why not worked? @ScaredOfMyRawr @Unsmart | Tech debt or anyone?
BIs there any way for batch processing, its fine. I want to implement that after every minute I add 50 users to 3rd party API likewise, I will be able to add all 400 users in 8 minutes. it will be fine... this schedules every day by cron trigger..is there any way I can do this. because by setTimeout or adding promise and fulfill it on 1 minute. both this approaches not working....
AIt’s not a timing thing, it’s a Workers runtime limit. On your current usage plan, you’re limited to 50 subrequests per request (when your scheduled task is run, it is considered one request). You can change your usage plan to standard, or unbound, and that will gice you 1,000 subrequests (e.g. HTTP calls) from your worker.
BI changed to standard, still same
ADid you deploy?
BYes deployed
AOnly thing I can guess at is maybe the subrequest limit is different for cron triggers, or perhaps it takes a little bit for your account to be fully migrated to standard. I would consider queues anyway though, as even if the limit is 1,000, you’re not far from that limit already.
Byes correct
B@ScaredOfMyRawr can you please explain some more about the queue approach, I am not familiar with cloudflare queue, was just reading the docs and got to know that, we have to make 1 producer worker for adding messages into queue and another consumer worker for fetching the message....
So I have to iterate over result array (400 customers) and add one by one into message queue, all this in producer queue...and in consumer queue, that logic and adding customer into klaviyo. right?
Can you please tell more about this approach
So I have to iterate over result array (400 customers) and add one by one into message queue, all this in producer queue...and in consumer queue, that logic and adding customer into klaviyo. right?
Can you please tell more about this approach
AYou got it spot on You already have your producer worker (your scheduled cron trigger) where you publish from, and then you can I think also just put a “queue” method on your worker and it will be your consumer too
You already have your producer worker (your scheduled cron trigger) where you publish from, and then you can I think also just put a “queue” method on your worker and it will be your consumer tooAJust put the customer ID or something into the message you put in the queue (or put in some JSON)
BOkay got it, thanks. Will try
BHello is it possible to add externals to the esbuild used by wrangler?
BAs far as I can tell in the documentation, there is not.
https://developers.cloudflare.com/workers/wrangler/configuration/#bundling
https://developers.cloudflare.com/workers/wrangler/configuration/#bundling
Use a
wrangler.toml
KNope
KThe only externals Wrangler currently has is
cloudflare:* and node:*
HWhat's the best way to set up service bindings with two workers, one written in Rust, the other in JS? Two separate project spaces referenced together via
wrangler.tomlYep
Rbtw, can anyone tell me that this post about? https://blog.cloudflare.com/sending-email-from-workers-with-mailchannels/
The Cloudflare Blog
MailChannels has created an email sending service specifically for Cloudflare Workers that removes all the friction associated with sending emails
RI am not quite understanding, is it just a promotion for mailchannels and in order to use it, I have to be a customer of mailchannels? let me know
UIt’s pretty much a promotion of mailchannels yes but you don’t need an account on mailchannels you just setup the dns records needed and send emails for free
RI really don't get it, let me clarify my misconception, how they can allow sending free email? is there any limit?
UThere’s some limits around how fast you can start sending out emails like you can’t randomly have a giant burst of thousands of emails can’t find the exact message rn from search but generally yeah you shouldn’t really have any issues using it fully free 
Rthat sounds too good to be true
RI have a provider that doesn't have api support and only exposes SMTP, so I was looking for some solution, but I think this would be enough
