Since we moved from AWS → self-hosting on Hetzner, our products have been much more stable/reliable. Our uptime is near-100% and we're never blocked from shipping. So the whole "self-hosting is fun until servers break at 2 AM" is a meme. You know which services go down all the... time? - Supabase - AWS - GitHub - Cloudflare Ask my competitors 😬 They would know!show more

8 myths about self-hosting. Let's kill them 👇

On Hetzner the outbound traffic is included too, so you skip the egress bill that quietly balloons on AWS.

Yep!

My website is even faster since i moved to vps

Same!

Same here. Hetzer, ansible, k3s. Own your infra

💪

I've been following your self-hosted "crusade" for a while. First of all, really well communicated and argued. Secondly, I'd genuinely love to hear your thoughts on what we're doing with Zerops, even though it's a bit of an antithesis to what you've been pushing. Would you care to feed this raw capability list (with an immediate roadmap) into whatever AI you're using, have a chat with it about it, and share your thoughts? 🙏 (happy to have a good old human chat as well anytime 🙂)

True. This is why we are building - Simple VPS/Cloud Servers on the infra that we built ourselves for our customers. Just few weeks away from the actual launch :)

Cool, but now you have all this DC stuff to monitor, act, open RMA requests, ensure your data survives when on site engineer makes a mistake, etc. So that meme “is a thing” 😉

Just did the same for our Staging env, going great so far with $40+ savings a month. Production is going to follow I guess!

Absolutely true, but scaling with AWS can be smoother for some startups.

Yes I comment on this in the full video too.

I only self host, 17 years now. To the point iv built my own base44.

besides, it’s way easier to run agents directly in production on live nginx, right? right???

Supabase went down for me today

Supabase goes down for someone everyday.

great lighting in the vid btw lol what’s your set up?

Haha thanks. It's very basic, a few key lights and a few side lights.

Self hosting is only bad if you have no idea how to do it. If you have a good grasp on infra you can save a lot using it. Also, everything can break at 2 AM, you can have backup servers ready to scale if that's something you need. And not all startups need that anyway.

Taking inspiration from your words, I am currently working up my self-hosting setup for my SaaS. Hope it will works fine 🤞🤞

Awesome! 💪

not needing permission from someone else's outage is the real win

Love me some CX43s I have several hetzner VPSes and the experience could not be better.

Self-hosting works until you need GPU clusters for inference. Does Hetzner even provision H100s on demand?

Unfortunately not 😒

Right, Hetzner is solid for web workloads. The moment you need H100s you're back on AWS or Azure.

@SimonHoiberg They're experimenting with inference:

@SimonHoiberg This confirms it. RTX PRO 6000 Blackwells, not H100s. Six hours to capacity, 10-minute TTFT at p99. Great for web workloads, not inference.

@hunvreus @SimonHoiberg Get your inference else where Jesus, does it all have to be in the same cloud provider for it to work ?!!! And there are a few inference providers who are far better than what was can give you. Try fireworks or Cerebras, etc...

@hunvreus @SimonHoiberg Exactly. You pick the right provider for the workload. Hetzner for web, Cerebras or Fireworks for inference. That was the argument from the start.

what did you end up using for user auth in your stack?

better auth

running multiple Kubernetes clusters in Hetzner. Some have over 4 years of uptime.

so true

It’s just the beginning But surprised to see supabase in the list Damn

I moved from supabase, vercel, cloudflare .. to hetzner using Dokploy. Everything live at one place. My servers, frontend, database, and redis. Speed gain is enormous.

I’m a big fan of self-hosting. It also is much harder to scale. DHH did it but they already spent millions. Cloudflare can scale from 0 to millions /w Workers. That’s just not something you can easily self host. But not everyone needs that of course!

@AniC_dev AI really helps with this I find. Along with the migration itself.

the big providers go down too, people just do not blame them the same way.

Absolutely

The ending of the video hits hard... 😰😰😰

Wow we're looking at migrating a DR site to hetzner and was worried about the uptime SLA thanks for the reassurance 👍

Effectivement... Je suis également avec Hetzner depuis des années dans ce format... Supabase existe en self-hosted et avec quelques tours de baguette magique il tourne merveilleusement bien en production. 😉

Hope you're not on GitHub. with your code.. :)

Dude hell yeah! And you haven't even mentioned the cost savings... 🫣

You’re a meme

This is a bit of a stretch. I am pretty sure you won't make a tweet when something you self host goes down