Live data from Hacker News

Migrating from AWS to Hetzner

digitalsociety.coop

181–190 of 643 posts

Re: Migrating from AWS to Hetzner

#181
post #143
post #137

Anyone here using https://github.com/vitobotta/hetzner-k3s Or https://github.com/kube-hetzner/terraform-hcloud-kube-hetzne... For a K3S cluster? Would love to hear any experience. Thanks!

We are using hetzner-k3s, super solid and easy to use. I love that it also utilizes the Hetzner Cloud Controllers to utilize the native Hetzner load balancers, volumes, network

Great, thanks!

Re: Migrating from AWS to Hetzner

#183
After being immersed in cloud-native hell for a few years, I'll say it:

This setup is probably also easier to reason about and easier to make secure than the messy garbage pushed by Amazon and other cloud providers.

People see Cloud providers with rose-colored glasses, but even something like RDS requires VPCs, subnets, route tables, security groups, Internet/NAT gateways, lots of IAM roles, and CloudWatch to be usable. And to make it properly secure (meaning: not just sharing the main DB password with the team) you need way more as well, and it's hard to orchestrate, it's not just an option in a CloudFormation script.

Sure securing a server is hard too, but people 1. actually share this info and 2. don't have illusions about it.

Re: Migrating from AWS to Hetzner

#184

I cannot overstate the performance improvement of deploying onto bare metal. We typically see a doubling of performance, as well as extremely predictable baseline performance. This is down to several things: - Latency - having your own local network, rather than sharing some larger datacenter network fabric, gives around of order of magnitude reduced latency - Caches – right-sizing a deployment for the underlying har…

Using the S3 API is like chopping onions, the more you do it, the faster you start crying.

Less to no crying when you use a sharp knive. Japanese chefs say: no wonder you are crying, you squash them.

Re: Migrating from AWS to Hetzner

#185

I cannot overstate the performance improvement of deploying onto bare metal. We typically see a doubling of performance, as well as extremely predictable baseline performance. This is down to several things: - Latency - having your own local network, rather than sharing some larger datacenter network fabric, gives around of order of magnitude reduced latency - Caches – right-sizing a deployment for the underlying har…

What is old is new again. My employer is so conservative and slow that they are forerunning this Local Cloud Edge Our Basement thing by just not doing anything.

"Do nothing, Win"

Re: Migrating from AWS to Hetzner

#186
There's a value curve for infrastructure, I'll use an analogy...

  Low Cost                                           High Cost
  ==============================================================
  FARM     WHOLESALER     GROCERY     RESTAURANT     DOORDASH    
  BUILD    CO-LOCATION    HETZNER     AWS            VERCEL           
While it's not a perfect analogy, in principle it holds true.

As such, it should come as no surprise that eating at a restaurant every day is going to be way more expensive.

Re: Migrating from AWS to Hetzner

#187
post #118

I think you can get much farther with dedicated servers. I run a couple of nodes on Hetzner. The performance you get from a dedicated machine even if it is a 3 year old machine that you can get on server auction is absolutely bonkers and cannot be compared to VMs. The thing is that most of the server hardware is focused towards high core count, low clock speed processors that optimize for I/O rather than compute. It…

I have seen hetzners cloud block storage to be quite slow. It became soon a bottleneck on our timescale databases. Now we're testing on netcup.com's "root servers" which are VPS with dedicated CPU cores and lots of very fast storage.

They limit them to 7500 IOPS, as stated in their docs. It also doesn't scale with size, the limit is there for every volume of any size.

Re: Migrating from AWS to Hetzner

#188

This is pretty bad still, with colocation you can get the costs down to 1/100th with good deals at datacenters especially ones that are struggling to attract customers. Most of your bill is power so if you rack efficiency optimized servers you can have a lot of compute for very cheap. In terms of networking many offer no-headache solutions with some kind of transit blend. I recently had to switch away from hetzner du…

But when you are Colocating you have higher upfront costs as you need to acquire hardware and also need to have somebody nearby the datacenter for hardware swaps in case of a failure, no?

You need to buy the hardware. However, you don't really need a dude on-hand to swap stuff on a daily basis, unless you're trying to host backblaze. The approach we take (with our data center 1000 miles away) is to provision excess machines. So if we need 6 machines we'll provision, say 8. Failure modes are always assumed to be "the whole machine" -- so a machine is either in service or not. Over time (years) one or two machines might fail in one way or another. Every couple of years we mount a rescue mission to repair/replace the bad machines, do some upgrades etc. We have redundant switches and routers and would make a special trip to replace one of those if there were a failure. The entire deployment has a "scaled to zero" cloud hot standby in place for the eventuality that the whole setup gets nuked somehow.

Re: Migrating from AWS to Hetzner

#189
post #91
post #28

Hetzner's ARM servers are the best kept secret in tech. Unbelievably capable and mindbogglingly cheap.

Have you encountered any software that wasn't compatible ?

Only a couple of times, but nothing that I use on production servers, only things that were very much hobby projects. For the sort of things we build an 8 core Hetzner ARM server outperforms an 8 core Digital Ocean x86 server by 10-20% for about a tenth of the cost.

Re: Migrating from AWS to Hetzner

#190

At my home, I have the PUD fiber Internet with Starlink as the backup WAN. I have two non mission critical servers we were using on AWS for and I set up two old laptops locally to run. Saving $1,000 a month now. I am looking at my $4,000 a month of mission critical servers now.

I began years ago hosting at home (because the business grew unexpectedly and initially the test server had been in my home). I wouldn't recommend it. Better to rent colocation space. The problem apart from having to provision reliable internet is you also need to provision reliable power, and cooling. And it gets noisy.
Post reply on HN