Serverless Is a Mistake: Stop Deploying to It!
ฝัง
- เผยแพร่เมื่อ 24 ก.ย. 2024
- Serverless is often hailed as the future of web deployment, but it might actually be slowing down your website. In this video, I discuss the drawbacks of serverless solutions like Vercel, Cloudflare, and AWS, and why they might not be the best choice for your projects.
🌍 My Website: www.janmarshal...
✅ Follow me on X: x.com/janmarsh...
📧 Business ONLY: jan@alenix.de
Yeah. Of course it always depends on the use case. But I think the pithy chunking I use is that there are not that many use cases for it. Pixel trackers for email opening, video transcoding. Auxiliary functions where the use is gone or expects to wait.
There have been many "anti" pushes over the years and I think they either have been overused or misguided from the start. Nosql was supposed to be "not only sql" but people thought it meant SQL-less but you cannot solve CAP so you still have SQL. No-ops was supposed to be about automation but no one could build the automation so you still have ops. Cloud was about removing traditional machines but it doesn't come with staff to run it. Devops was supposed to remove friction but we couldn't make the dev&ops people do the extra config work so this idea has moved to IDPs (time will tell what will happen next). Dynamic languages skipped on types but now they are coming back as Typescript/Sorbet and Python types.
These are huge topics with lots of nuance that I probably could not explain even if I had a huge comment box but I think just the three-tier app is not bad to start with if I don't know anything about the project or the future.
WOW, Chris. I never really looked at it from this perspective, but you're absolutely right. It’s fascinating to see how these ideas evolve over time, and you explained it so well.
When Jan posts, i watch
Can you make a micro service app using the typescript and docker.
Have you ever heard of Serverless first thinking?