Is anyone using workers to proxy websockets getting 502 errors currently? Suddenly all our websocket
Is anyone using workers to proxy websockets getting 502 errors currently? Suddenly all our websockets, across multiple accounts, stopped working
WHey, got a repro URL? and is this fully client <-> Worker or is there an origin involved?
LOrigin, normally connects to a cloudflare LB, but I have now tried arbitrary origins that support wss and it all returns 502. It also swallows our logs so its like something is broken in the upgrade path
You can try making a wss connection to here `atlas-mainnet.helius-rpc.com?
You can try making a wss connection to here `atlas-mainnet.helius-rpc.com?
Worigin is definitely up and working?
L100%, I can connect to it. I also have multiple origins across accounts that all work going to directly
Land also tried putting origins that I don't own that support wss in there and same error
Wi see you have an escalation open internally already
WIt's a p1 escalation, I'm tracking it now
WWe've reverted the release causing this, thanks for the report and sorry for the issues @liam | helius.dev !
Lnp, ty for the help!
Ahi. is this still valid?
https://blog.cloudflare.com/sending-email-from-workers-with-mailchannels
Getting:
Thanks
https://blog.cloudflare.com/sending-email-from-workers-with-mailchannels
Getting:
Thanks
The Cloudflare Blog
MailChannels has created an email sending service specifically for Cloudflare Workers that removes all the friction associated with sending emails.
UDid you go to the link the error said? Should explain what you need there
Chttps://developers.cloudflare.com/workers/testing
what are service workers in this context? can I test regular module workers with vitest pool?
what are service workers in this context? can I test regular module workers with vitest pool?
Uservice workers are the old form of creating workers, module workers are the new form
Uhttps://developers.cloudflare.com/workers/reference/migrate-to-module-workers/#migrate-from-service-workers-to-es-modules is the migration guide and can show the differences from them
JSadly not a direct solution to your problem, but I would recommend you avoid MailChannels. Cloudflare's continued partnership with them is very disappointing after their security issues and terrible response. See https://canary.discord.com/channels/595317990191398933/779390076219686943/1226529017763463219 for more details.
Athanks. yup, went to the link, talks a lot about accounts, thought I needed to create an account with MailChannels. tried adding cfid for worker and now it's returning
null
Ainteresting, thanks
Whow long can a websocket connection be open on free plan
Cwhen i create a new file in my javascript worker and deploy it the file is still not saved
CIt doesn't save the new javascript file
II don't think there's a limit.
LI'm thinking about migrating a Nuxt application from Vercel to CF pages, but I'm curious if I'll be able to use the cache JavaScript api on the Nuxt server side code?
To be clear, I'm talking about this: https://developers.cloudflare.com/workers/runtime-apis/cache/
clearly works within vanilla workers just wondering what the limitations are with SSR frameworks like Nuxt on pages
To be clear, I'm talking about this: https://developers.cloudflare.com/workers/runtime-apis/cache/
clearly works within vanilla workers just wondering what the limitations are with SSR frameworks like Nuxt on pages
Uyou can use cache, pages ssr functions are just workers
Lawesome ty
Lwhere is durable objects channel?
H#durable-objects
JI have a cron task that should add jobs to a queue. In order to avoid a duplicate jobs being added to the queue the next time the cron runs, I need to keep track of which jobs have been triggered. What's the best way to do that? My initial thought was KV, but I don't think it has any support for something like a transaction.. and neither does D1?
JThe enqueuing and the storing of the "this job has been enqueued" record needs to be atomic
Mwith the new RPC calling, should you still have zod schemas passing the incoming parameters?
UI mean you can if you want. You didnt even need zod schema when you didnt have RPC its use is mostly do you trust that you actually sent the correct data format or do you want to double check it. I personally never used it because I trusted types
Mhow were you able to avoid using it pre-RPC? e.g. let's say the payload is in the body, and you do
let body = await req.json() then the typeof body would be "unknown"Uas <type>
Mah ok, i guess as long as you were wiring up your calls correctly then the parser isn't needed
my thoughts are more along the lines of deploying new versions of the worker, what if you change the payload? you're going to probably start getting some cryptic errors without a front-loaded parser
my thoughts are more along the lines of deploying new versions of the worker, what if you change the payload? you're going to probably start getting some cryptic errors without a front-loaded parser
GHi, quick question
If i install a package such as
Currently I cannot see any logs or output when trying to use it
If i install a package such as
aws-sdk/client-s3 in a typescript worker, can i expect it to just work as normal in a cloudflare worker? Currently I cannot see any logs or output when trying to use it
Wwow awesome
FIs it viable to use only CF workers for the backend of my mobile app? I can't find any drawbacks some maybe someone could share some experiences. I would greatly appreciate that
ISo there's not, but there is this: https://developers.cloudflare.com/network/websockets/#technical-note
IWhen Cloudflare releases new code to its global network, we may restart servers, which terminates WebSockets connections.
ISo you definitely want to implement retries of some sort on the client.
Wnot a limit on duration but note https://developers.cloudflare.com/network/websockets/#what-happens-if-my-site-exceeds-the-number-of-concurrent-websockets-connections-that-cloudflare-expects for concurrency
Chow do I pass
it's telling me
env here? https://developers.cloudflare.com/workers/testing/vitest-integration/get-started/write-your-first-test/#write-testsit's telling me
env is missing properties from my Environment interface
CI'll try using javascript instead of typescript I guess
BI've been trying to get my www record to work but it doesn't. I'm using bulk redirect rule to redirect to apex but in the browser i keep getting "server IP address could not be found." i'm using 192.0.2.1 for A record and 100:: for AAAA record. my apex domain works on its own though.
Csounds like dns cache, don't get what that has to do with workers/this channel though
Bwhoops wrong channel
Banyways, i tried it in incognito tab and on a different device but it's still not working for some reason
Bnot sure if i messed up a config somewhere or what
Ccan you share the domain name in #general-discussions or a thread in #general-help?
(entirely possible you are using your default dns resolver which is usually your ISP's, that would also cache across devices)
(entirely possible you are using your default dns resolver which is usually your ISP's, that would also cache across devices)
BNVM you were right. i used a VPN and it works now
