it makes me sad that part of logpush's backend seems to be written in Go instead of TypeScript 😦

Workers/http should be at the same rate, but just way less request per batch

yea why would you want typescript for that? I'd assume rust over Go

because then it can run in a Worker!

logpush itself isn't a workers service I imagine, iirc it predates them, also just the "type" of service it is, needing to batch/be very efficient/etc, probably not the best use case

@Erisa | Support Engineer @Chaika thanks for the input. sounds like i should hold off and maybe hope that tunnels find a way for internally supporting tcp 😄

No problemo, would be nice indeed if Workers could access tunneled tcp directly 🙂

looks like they're just talking about using cloudflare access cli for accessing a database

which under the hood just uses the same tcp over websocket stuff iirc

Go ❤️

indeed.. lots of 'almost' solutions

Why can't they? The orange cloud block or?

even with a limitation of only tunneling to your own workers would be nice. 🙂

lol it's what browser rendered ssh uses as well, tcp over websocket parsing locally. It works great, but is a bit funny to me

simplifies audits anyhow

Same reason specturm cant, cf tunnels dont expose TCP in a way that can be consumed by anything other than over https with the websocket proxy thing

ah fair

Workers can connect over the websocket proxy which I mentioned but they didnt seem to want to do that

well more like I'm not sure how to get the postgresql library to use it 😛

telling a worker to participate in a specific (tunnel?) would be cool

sounds.. heavy tho

that would be incredibly amusing tho.. cloudflare made those changes to add raw sockets to workers, and then we just convert them back to websockets. 😛

Ig cloudflare needs an update to it after almost 10 years x3

wrangler routes setting I have a problem with this #799437470004412476

Hey everyone, what would be the right place to ask for help with Cloudflare Images?

This channel, #1104040620948992090 or on the community forums ()

Ok, well I was basically trying to use Sourcing Kit to import images, but I wanted to point it to an R2 bucket. Now R2 does support an S3 API, but I couldn't get it to work, it keeps telling me my credentials are wrong. Is there a workaround I'm missing, or is this not possible to do yet?

Turns out I just had to wait for the SRV record to propagate, now it works just fine :D

Ah, awesome! 😄