[HN Gopher] A day at Hetzner Online in the Falkenstein data center
       ___________________________________________________________________
        
       A day at Hetzner Online in the Falkenstein data center
        
       https://www.youtube.com/watch?v=gIjdKIMQh4s
        
       Author : speckx
       Score  : 154 points
       Date   : 2025-11-18 16:18 UTC (6 hours ago)
        
 (HTM) web link (www.igorslab.de)
 (TXT) w3m dump (www.igorslab.de)
        
       | Freak_NL wrote:
       | Location of Hetzner's Falkenstein data centre, for the curious:
       | 
       | https://www.openstreetmap.org/way/606528332#map=13/50.46220/...
        
       | boldlybold wrote:
       | Video didn't display for me even after turning off content
       | blockers, here's the url:
       | https://www.youtube.com/watch?v=gIjdKIMQh4s
        
         | tcfhgj wrote:
         | works with Firefox + uBlock Origin + canvas blocker
        
       | ArtTimeInvestor wrote:
       | Hetzner is awesome.
       | 
       | I wish they would go public.
       | 
       | Would be very interesting to see how their business is doing
       | compared to IONOS and OVH.
       | 
       | We need more public cloud companies in Europe.
        
         | progbits wrote:
         | Why would you want a company you like to go public? To keep
         | chasing profits to the detriment of everything else?
        
         | moooo99 wrote:
         | Why in the world would one want a company like this to go
         | public? They are a very stable, established and profitable
         | private entity. Based on what Hetzner is doing right now, it
         | seems like the current way of operating that is intended by the
         | leadership is closely aligned with what their customers want.
         | This is often the first thing that goes out the window once a
         | company becomes public
        
         | Freak_NL wrote:
         | > I wish they would go public.
         | 
         | I wish that they'll just keep on running a good profitable
         | business without going public, or getting bought by Amazon, or
         | otherwise shifting focus to providing shareholder value.
        
           | sdoering wrote:
           | This. I am a happy paying customer. Recommend them to my
           | clients whenever it makes sense (more often than not). And
           | hope they just keep running like they do as profitable
           | business, like you said.
        
         | philipwhiuk wrote:
         | > I wish they would go public.
         | 
         | If they go public they will be bought by foreign private
         | equity.
        
         | baxtr wrote:
         | Yes I agree, we need more public companies in Europe.
         | 
         | Private companies are inherently less social since they don't
         | allow ordinary people to participate in growth.
         | 
         | In this sense, they're selfish.
         | 
         | PS: yes I know that there are also downsides to public
         | companies. But looking at the trade-offs I prefer that success
         | can be shared as broadly as possible.
        
           | bell-cot wrote:
           | At least in America, "a successful private company went
           | public" often translates into "ordinary people got a bit of
           | gold, selfish vulture capitalists butchered the goose, and
           | there was precious little success or growth for anyone after
           | that".
           | 
           | (Also - might your "allow ordinary people to participate"
           | sympathies extend to people who would like to participate in
           | your own financial affairs?)
        
             | baxtr wrote:
             | Yes, exactly!
             | 
             | Selfish VC becoming filthy rich through an IPO is exactly
             | my point. Up to an IPO a private company will only make
             | their owners rich - in your example "selfish vulture
             | capitalists".
             | 
             | After an IPO anyone can participate. When Google, Amazon,
             | Apple went public, VCs got rich. Everyone after that
             | included every day people like you and me.
        
               | bell-cot wrote:
               | No - the selfish _vulture_ capitalists are the outsiders
               | who purchase a private company which has been successful
               | for many years, then butcher it. There is no IPO - it is
               | "you own X, and we are offering you $Y million to sell it
               | to us".
               | 
               | After that - X's best assets are sold off (the VC's get
               | the money), X goes deeply into debt (again, the VC's get
               | the money), many of the employees are laid off, and X
               | generally goes bankrupt within 7 years - because what is
               | left of it can't make the payments on the debt.
        
               | Gooblebrai wrote:
               | I don't understand your comment. You are both talking
               | about public companies, and suddenly you are now talking
               | about private equity?
        
               | bell-cot wrote:
               | (You're right - I made a mess of things, and inter-
               | mingled the cases where a privately-owned company is sold
               | directly to private equity / vulture capitalists, and the
               | case where a privately-owned company "goes public" - but
               | that still does not lead to a happy ending.)
        
               | lotsofpulp wrote:
               | >the selfish vulture capitalists are the outsiders who
               | purchase a private company which has been successful for
               | many years, then butcher it.
               | 
               | Sounds like the selfish vulture capitalists are the
               | insiders who sell the company.
               | 
               | >X's best assets are sold off (the VC's get the money), X
               | goes deeply into debt (again, the VC's get the money),
               | many of the employees are laid off, and X generally goes
               | bankrupt within 7 years - because what is left of it
               | can't make the payments on the debt.
               | 
               | This doesn't make any sense, because X is the original
               | asset. If part of X is sold, then the remaining portion
               | of X loses value (assuming the sold part is the good
               | part). If X is used as collateral, then it also loses
               | value.
        
           | nchmy wrote:
           | they offer by far the best value servers in the world. If
           | that's selfish, so be it
        
         | jeffbee wrote:
         | The comparison with OVH is a good one. For some reason OVH has
         | much worse PUE, self-reporting 1.24 vs. Hetzner's 1.11.
         | Operating costs are basically just electricity for these
         | places, so their margins are that much worse.
         | 
         | For further comparison, Google at a similar latitude in Saint-
         | Ghislain, Belgium, claims 1.08.
        
           | arcanemachiner wrote:
           | I don't understand anything in your comment, so I'm
           | shamelessly posting some info from an LLM:
           | 
           | > PUE (Power Usage Effectiveness) measures datacenter
           | efficiency - it's total facility power divided by IT
           | equipment power. A perfect 1.0 means all power goes to
           | servers; higher numbers mean more waste on cooling/overhead.
           | OVH's 1.24 vs Hetzner's 1.11 means OVH burns 24% extra power
           | on non-IT stuff, hurting margins since electricity is their
           | main cost. Google hits 1.08 at similar latitude.
        
         | moffkalast wrote:
         | Damn Americans, can't think of anything other than "I wish I
         | could make a buck off this, idgaf if it gets destroyed in the
         | process".
        
         | Anarch157a wrote:
         | The last thing we need is more enshitification on this space.
         | 
         | I just migrated my selfhosted email server to Hetzner and I
         | don't want them turning into monstrosity like AWS or Azure,
         | with theit miriad of ways to nickel and dime the customers.
        
         | zelphirkalt wrote:
         | Compared to IONOS, it is day and night. IONOS in my experience
         | is just worse in every category that matters. IONOS has buggy
         | and bad wannabe SPA web frontend, broken OS images, bottlenecks
         | in their provisioning API, that they don't care to fix, and 10x
         | the costs.
         | 
         | If you are not desperate for one specific offering of IONOS,
         | that Hetzner doesn't offer, and you have any other options,
         | there is no good and justifiable reason to go for IONOS. IONOS
         | is a low effort, low quality, high marketing spend shop.
         | Hetzner is almost the opposite. They don't waste money on
         | advertising, they cost much less to rent servers from, their OS
         | images work, they got good technical support.
        
       | esafak wrote:
       | Do tech companies like Hetzner count as _mittelstand_ s?
        
         | IlikeKitties wrote:
         | "Mittelstand" in Germany seems to be roughly: below 500
         | Employees, below 50 Million EUR turnover. Hetzner makes more
         | like 400 Million EUR turnover but with 300 Employees. So it's
         | technically above it but they are propably a very tight margin
         | buisness. As such i'd say yes.
        
         | elcapitan wrote:
         | Mittelstand is singular (like middle class).
         | 
         | I would say decidedly yes for Hetzner. Mittelstand can somewhat
         | be characterized by size, but it's mostly that they are more
         | like a family, have different values and more long term
         | thinking than the larger industry and public companies. It's
         | also kind of a brand that many companies would like to
         | attribute to themselves, even much larger companies that like
         | that seal of quality.
        
       | noir_lord wrote:
       | Had a server in that DC for a few years until recently (no knock
       | on Hetzner they where excellent, just didn't need it any more).
       | 
       | Robot is a brutally functional tool but it does function (well)
       | and had zero issues.
        
         | john01dav wrote:
         | Regarding Robot, I think that it's completely fine for what it
         | is. I almost never interact with it, and instead just configure
         | my server as I see fit over ssh. Hetzner's value proposition is
         | extremely cheap no frills servers -- you're paying for the
         | server, not the management interface. If you want management
         | interfaces that do a lot of useful work, use a cloud.
        
           | noir_lord wrote:
           | Yup, which is why I said brutally functional and not broken.
           | 
           | I _like_ brutally functional it just needs to get me to SSH
           | and I 'm good.
        
       | jsheard wrote:
       | Der8auer did some similar videos a few years ago, but in English:
       | 
       | https://www.youtube.com/watch?v=5eo8nz_niiM
       | 
       | https://www.youtube.com/watch?v=V2P8mjWRqpk
       | 
       | The second is especially interesting as it demonstrates Hetzner's
       | unique semi-custom hardware.
        
       | yrro wrote:
       | I love Hetzner. That said their IPv6 support is poor. A server
       | gets a /64 only, if you want a /56 (allowing 256 container
       | networks) then you have to pay EUR15. As for virtual networks:
       | they only support IPv4!
       | 
       | At least they're not as bad as Azure... :)
        
         | matt-p wrote:
         | Why do you need ipv6 on your internal network? Is 10/8 really
         | not enough/overlap? For 99.99% of people it's fine for the
         | internal interfaces and if anything actually simplifies
         | configuration.
        
           | yrro wrote:
           | The purpose of a network is to allow any two consenting
           | parties to communicate. IPv4 cannot deliver that if either
           | party has an RFC1918 address. NAT is a foul perversion of
           | this foundational principle of the Internet Protocol.
        
             | matt-p wrote:
             | On your *internal* network e.g the thing between your
             | postgres VM and your webserver (or whatever). Not arguing
             | against it on the public/wan connection.
        
               | yrro wrote:
               | There is no such thing as an 'internal' network.
        
           | fulafel wrote:
           | For a lot of use cases a major advantage of IPv6 is to get
           | away from ambiguous rfc1918 addressing.
           | 
           | You can then just put an allow rule between arbitrary v6
           | addresses anywhere on the internet when you need connectivity
           | without any other hacks like proxies, NAT, etc and the
           | associated complexity and addressing ambiguity/context
           | dependence of rfc1918 addresses.
           | 
           | So fex you can just curl or ssh to your
           | mycontainer.mydomain.net or you can put an allow rule from
           | mycontainer.mydomain.net to a vm or laptop on your home
           | network.
           | 
           | Internetworking, they call it.
        
             | matt-p wrote:
             | I'm talking about an _internal_ network, not the public
             | connection.
        
               | fulafel wrote:
               | The context in the GP comment was generally getting v6
               | connectivity for containers.
               | 
               | "Internal" is a context dependent term that you
               | introduced. But to give a use case for that, for example
               | you might want to have (maybe at a future date) two hosts
               | on your networks on AWS and Hetzner talk to each other,
               | still without allowing public connectivity.
        
           | mulmen wrote:
           | The benefit of IPv6 is that I don't need an "internal"
           | network. Everything is on the Internet and the firewall is
           | the only thing that gets in the way.
        
         | speedgoose wrote:
         | I disable IPv6 and I'm somewhat scared of the concept of having
         | containers with public IPs.
        
           | yrro wrote:
           | Routable != routed.
           | 
           | If your containers have a Global Unicast Address then it's
           | possible to look at connetion logs and figure out which
           | container made a particular request, for instance.
        
             | matt-p wrote:
             | Yes, not addressable is even safer. Especially so for
             | someone not specialized in networking.
        
               | yrro wrote:
               | It's not safer: it impedes observability.
        
               | matt-p wrote:
               | It doesn't impeed observability for goodness sakes. It
               | does however impeed accidentally opening up your internal
               | network because you don't really understand your
               | firewall/virtual router/whatever.
        
               | yrro wrote:
               | Of course it impedes observability. With IPv6, I can see
               | the IP addresses of the containers that connect to a
               | service. With IPv4, I get (at best) the IP address of the
               | container host, thanks to NAT.
               | 
               | Are you also afraid of port forwarding? Have you
               | considered that your ISP could choose to send your router
               | packets destined for RFC1918 addresses?
        
         | bonyt wrote:
         | Does each container network of the 256 really need its own /64?
         | Is there some constraint that doesn't let them work on a /72?
        
           | matt-p wrote:
           | A container should absolutely not even need a /72. The
           | traditional reason for /64 is for slaac but you most
           | certainly don't need that for one container (if at all
           | honestly).
        
             | yrro wrote:
             | Indeed, a host should be able to request a /64 via
             | DHCPv6-PD and split that between millions of container
             | networks. But you can't do that on Hetzner (or anywhere
             | else).
        
               | matt-p wrote:
               | Yeah that obviously only works on /56 and above because
               | networks should be a minimum of /64. I use k3s and each
               | host has a /64; cilium just gives each pod a /80 and the
               | host does NDP and stuff. Works fine, no need to require
               | dhcp6.
        
           | yrro wrote:
           | In practice this can be made to work but a networking expert
           | can probably explain better than me why splitting a prefix
           | into chunks smaller than a /64, and assigning them to virtual
           | networks within a host is a bad idea.
           | 
           | In Hetzner's specific case: they won't give me one or more
           | additional /72s: only a /56 if I pay for it. Per server.
        
       | colonCapitalDee wrote:
       | I have a tiny Hetzner VPS (2 vCPUs, 2 Gb RAM) in their west us
       | datacenter that costs me $5.59 a month. I get 1 TB a month free
       | outgoing bandwidth and unlimited incoming bandwidth, plus
       | additional outgoing bandwidth at a rate of $1.20 per TB. I host
       | my personal project's git LFS server there, a file server, and a
       | Caddy instance that proxies over Tailscale to a more powerful box
       | in my apartment. It's a great homelab architecture and I couldn't
       | be happier with it. Thanks Hetzner!
        
         | kalaksi wrote:
         | They still have VPSes with 2G of RAM? I'm checking the cloud
         | price page* and you can get CX23 with 2 vCPU, 4G RAM and 20TB
         | of traffic (seems to say that it's 1TB for US) for
         | 3,49EUR/month (~4 USD).
         | 
         | You can save additional 0,50EUR if you go with only IPv6.
         | 
         | Maybe it's location dependent, I can't get it to show me prices
         | for US.
         | 
         | * https://www.hetzner.com/cloud/
        
           | stavros wrote:
           | I think they retired a bunch of machines one or two weeks
           | ago. I was prompted to upgrade my machine at no extra cost,
           | anyway.
        
           | colonCapitalDee wrote:
           | Yep, just checked, 2 GB RAM. That CX23 sounds like a great
           | deal, 20 TB of free outgoing is ridiculous. But I live in
           | west us and I rarely hit my 1 TB free bandwidth cap anyway,
           | so the added latency isn't worth it
        
       | tcdent wrote:
       | Hetzner certainly has this cult-like following mostly because of
       | their low cost.
       | 
       | I assume it is a recent push toward these kind of open frame,
       | super minimalist, consumer hardware based systems (I don't speak
       | german and didn't translate the video).
       | 
       | It looks like they're using lots of consumer hardware and very
       | little redundancy; you'll notice that the power supplies are
       | generic ATX units and they're not doubled up. And then they're
       | also running the onboard networking with a second connection
       | which looks like it's for just a management system. Might not
       | even be 10 gigabit networking.
       | 
       | It's interesting that in an era where almost all of the major
       | players are moving toward cable-free arrangements i.e. backplanes
       | with fully integrated power and networking, etc., they're instead
       | opting for the rat's nest of cabling. It must have something to
       | do with lower labor costs vs hardware costs. The amount of
       | density that they are achieving with those systems is also
       | incredibly low relatively speaking.
        
         | radiator wrote:
         | 25 Gigabit networking is mentioned in the video. As for the low
         | density, it might have something to do with their mentioned
         | freecooling concept, which does not use cooling machines, but a
         | controlled stream of air coming from outside.
        
           | lifty wrote:
           | This summer their Amsterdam DC went down because of the
           | record breaking temperature in Amsterdam. Wondering if they
           | use the same cooling technique there. They did mention it was
           | because of cooling.
        
             | arend321 wrote:
             | Hetzner does not have a datacenter in Amsterdam
             | 
             | https://docs.hetzner.com/cloud/general/locations/
        
               | lifty wrote:
               | Ups, my bad. Confused the incident, it was Scaleway.
        
         | g-mork wrote:
         | Same old world thinking.. Google use single PSUs too, real
         | redundancy came from having multiple machines, and Hetzner
         | certainly makes that cheap enough to accomplish on a budget.
         | You can also pay for 10 Gbit as an option with Hetzner, and a
         | bunch of other custom upgrades, but the further you move
         | outside their sweet spot the more it's going to cost.
        
         | showerst wrote:
         | I think a lot of the following comes the old-fashioned nature.
         | In an age of hyperscalers that want to sell you a private cloud
         | and full stack certification to make it your job to use their
         | software and want to charge you $500/mo for a server with 64GB
         | of ram and decent bandwidth, Hetzner will just rent you a
         | server, relatively cheap.
         | 
         | Amazon has _95 pages_ of EC2 instance types. They have so many
         | products that I literally had to google the name to know what
         | product type to put into the estimator to get a boring server.
         | 
         | What's all that data center best practice get you[1], the
         | customer, if it doesn't provide lower prices and higher
         | availability?
         | 
         | [1] -- I'm assuming you are not Netflix. After some scale all
         | those crazy AWS services are pretty great to have.
        
         | TiredOfLife wrote:
         | > I assume it is a recent push toward these kind of open frame,
         | super minimalist, consumer hardware based systems
         | 
         | If by recent you mean that was how Hetzner has been since start
         | (20+ years ago) then yes
        
         | littlecranky67 wrote:
         | They no longer use consumer grade off-the-shelf hardware. They
         | use their own mainboards (yes, custom layouts due to their
         | unique cooling solution). Checkout the video from Der8auer on
         | youtube.
        
         | zelphirkalt wrote:
         | Their technical support is also quite good, at least in my
         | experience. Any questions I had they easily answered in a way
         | that was easy to understand for me as a technical person. Not
         | sure how it is for a non-technical person. But they might not
         | rent Hetzner.
        
       ___________________________________________________________________
       (page generated 2025-11-18 23:01 UTC)