4 ms·

I'm sure there are much bigger and more worthwhile criticisms to be had than this. It's something they should fix and if they did would you suddenly switch to

by s_dev 8mo ago

I'm sure there are much bigger and more worthwhile criticisms to be had than this. It's something they should fix and if they did would you suddenly switch to Scaleway? I think you would consider other factors first. A good critique for example is OVH lost a lot of customer data due to a fire. Where was the redundancy? That would make me think twice before switching to OVH.

troupo 8mo ago
> It's something they should fix and if they did would you suddenly switch to Scaleway? You know why I have this screenshot? Because I literally tried to switch to "great European alternative" that is "as slick as DO". After a third or a fourth screen, most of which felt completely isolated and disconnected from any previous ones, I gave up on the screen that couldn't handle a standard European address. This was literally the point that I gave up. So I went ahead... and signed up with Hetzner. Edit So I decided to try again. Literally the first page of account sign in tried to trick you into accepting tracking Since I apparently had an account, I could login... So redirected to a subdomain with the same cookie popup. On a site that is solely for billing address collection which then redirects you to a third domain with the a similar but different popup. Which ends up on an empty page indistinguishable in "usability" from Hetzner (or worse) That's the end of my experience of my "European DO that is Scaleway". They did fix the addresss boxes, kudos to them
alberto-m 8mo ago
A lack of Unicode support in 2026 is like someone coming with dirty clothes to a job interview: it might not affect too much how the work is done, but immediately raises doubts about the underlying level of professionalism.
- troupo 8mo ago
  That was 2024 (still inexcusable), they managed to fix it at one point https://news.ycombinator.com/item?id=46734771 https://news.ycombinator.com/item?id=46734771
u8080 8mo ago
Hetzner/Linode were MITMing their client(jabber.ru): https://notes.valdikss.org.ru/jabber.ru-mitm/ https://notes.valdikss.org.ru/jabber.ru-mitm/
- troupo 8mo ago
  Was it Hetzner, or was it an attacker hosting on Hetzner/Linode?
  - direwolf20 8mo ago
    It was the German equivalent of the NSA, with the German equivalent of a National Security Letter, sent to Hetzner to force them to intercept this customer's traffic. The same thing happens in the USA.
    - troupo 8mo ago
      Sigh Time to setup my own dedicated servers.
      - direwolf20 8mo ago
        The German NSA seemed unable to access the server as they only intercepted the traffic. They got a TLS certificate from Let's Encrypt by intercepting traffic. If the app had used public key pinning, and the server had full disk encryption, this wouldn't have been enough for a compromise.
celsoazevedo 8mo ago
> A good critique for example is OVH lost a lot of customer data due to a fire. Where was the redundancy? That would make me think twice before switching to OVH. I lost a VPS in that fire, but I was up and running a few hours later with a new VPS at a different OVH location. Not to deflect blame away from OVH and their large screw up, but we should never rely only on the redundancy of the hosting provider. Even on AWS, I wouldn't trust them to not lose my data if one of their datacenters burns down. At the time I was making regular backups to two different providers with servers somewhere else. When I noticed that it was serious, I ordered a new VPS and restored everything. If OVH itself went down, I could have used Scaleway, Hetzner, Contabo, etc.

by s_dev 8mo ago

troupo 8mo ago

> It's something they should fix and if they did would you suddenly switch to Scaleway? You know why I have this screenshot? Because I literally tried to switch to "great European alternative" that is "as slick as DO". After a third or a fourth screen, most of which felt completely isolated and disconnected from any previous ones, I gave up on the screen that couldn't handle a standard European address. This was literally the point that I gave up. So I went ahead... and signed up with Hetzner. Edit So I decided to try again. Literally the first page of account sign in tried to trick you into accepting tracking Since I apparently had an account, I could login... So redirected to a subdomain with the same cookie popup. On a site that is solely for billing address collection which then redirects you to a third domain with the a similar but different popup. Which ends up on an empty page indistinguishable in "usability" from Hetzner (or worse) That's the end of my experience of my "European DO that is Scaleway". They did fix the addresss boxes, kudos to them

alberto-m 8mo ago

A lack of Unicode support in 2026 is like someone coming with dirty clothes to a job interview: it might not affect too much how the work is done, but immediately raises doubts about the underlying level of professionalism.

troupo 8mo ago
That was 2024 (still inexcusable), they managed to fix it at one point https://news.ycombinator.com/item?id=46734771 https://news.ycombinator.com/item?id=46734771

u8080 8mo ago

Hetzner/Linode were MITMing their client(jabber.ru): https://notes.valdikss.org.ru/jabber.ru-mitm/ https://notes.valdikss.org.ru/jabber.ru-mitm/

troupo 8mo ago
Was it Hetzner, or was it an attacker hosting on Hetzner/Linode?
- direwolf20 8mo ago
  It was the German equivalent of the NSA, with the German equivalent of a National Security Letter, sent to Hetzner to force them to intercept this customer's traffic. The same thing happens in the USA.
  - troupo 8mo ago
    Sigh Time to setup my own dedicated servers.
    - direwolf20 8mo ago
      The German NSA seemed unable to access the server as they only intercepted the traffic. They got a TLS certificate from Let's Encrypt by intercepting traffic. If the app had used public key pinning, and the server had full disk encryption, this wouldn't have been enough for a compromise.

celsoazevedo 8mo ago

> A good critique for example is OVH lost a lot of customer data due to a fire. Where was the redundancy? That would make me think twice before switching to OVH. I lost a VPS in that fire, but I was up and running a few hours later with a new VPS at a different OVH location. Not to deflect blame away from OVH and their large screw up, but we should never rely only on the redundancy of the hosting provider. Even on AWS, I wouldn't trust them to not lose my data if one of their datacenters burns down. At the time I was making regular backups to two different providers with servers somewhere else. When I noticed that it was serious, I ordered a new VPS and restored everything. If OVH itself went down, I could have used Scaleway, Hetzner, Contabo, etc.