Always nice to see some balance in a post but this is definitely an anti-serverless post 😂 And that’s fine, you didn’t have a great experience, own it and enjoy your rant! 🍿

Hard agree on this though! Still plenty of work to be done on local DX

> This led us to build an elaborate and overly complex pipeline… for analytics events

On this point, why are analytics events in your hot path when aiming for sub 10ms? 🤔

> limitations imposed by the serverless model

Lots of teams believe the hype and expect serverless to be a silver bullet. When has any technology with that promise lived up to expectations?

You simply can’t get the incredible gains of serverless without losing some features you’re used to

All that said, the two key limits @unkey.com hit are clear reasons to avoid serverless. But if these are the USP of your product why were they not considered upfront?

> when you're trying to build a sub-10ms API, it's a showstopper
> customers couldn't self-host Unkey

> Our metrics became elaborate JSON logs

One word. CloudWatch

this is also why serverless != compute. Serverless is an ecosystem of managed primitives, not just cheap ephemeral compute!

> We found ourselves constantly evaluating and integrating new SaaS products

This is the tradeoff between a nimble, cutting-edge cloud and a mature, full-featured cloud. Many of the services they are reaching for are simply baked into the AWS experience…

And in actual fact they have moved to AWS Fargate (which is - at least marketed - as a serverless service!) 😅

First off, after reading thru the post it’s clear it should be titled Why We Are Leaving Cloudflare. Most of the constraints are Cf-specific, and wouldn’t be an issue on AWS

fetch?

This sounds cool Ben! Would love to see what this looks like. I understood Score to be focused on container platforms?

Just came across this, looks interesting medium.com/air-pipe-blo.... Uses @cloudflare.social Pingora