Comment by Filligree
3 years ago
I ran a server there for years. Eventually it developed a fault on... either the PCIe bus, one of the paired NVMes, or something in that vein. This manifested as poor performance, followed by the NVMe falling off the bus, often followed by the machine rebooting. Upon doing so it would stall at the BIOS, complaining that it had been overclocked. It had not been overclocked... as far as I know, at any rate.
I went back and forth on this with support for several months. They were not able to reproduce (it was a rare fault, happening on average only about once every few weeks), and blamed me for 'overclocking'. Eventually I shut off the account.
Hetzner is cheap, but you're on your own if you need any form of help.
I had a similar problem, where my AMD server kept freezing up about once a month. It was my critical mail server, and used by other people, so not good.
Unfortunately I just kept reacting to it by rebooting, instead of writing to Hetzner support.
When I did eventually write to them, they scheduled a date for complete replacement of the main hardware transferring over the SSDs without any fuss, and the problems have never reoccurred, so that was great.
The reason I didn't write to them for a long time is I'd heard stories like yours, and I thought I'd have to gather lots of evidence to prove the server was unreliable. I did that and presented it, but I was still surprised at how instantly they just proposed a date for replacement without any discussion.