News for hackers. Stuff that matters.
Re:Show HN: We moved from AWS to Hetzner, saved 90%, kept IS... by OutOfHere on Saturday June 21, 2025 @01:01PM (#44339035)
Hetzner's biggest problem is that they can and do terminate a user's account without warning if the user starts using CPU resources very heavily or for any reason. This is for very legal usage, of course. This can and does happen to people in months. When this happens, consider your data lost and your account blocked. They will offer no explanation whatsoever, and will even send you a bill for the full month. Hetzner simply cannot be trusted, not even a little bit.

As for OVH, they don't do the above, but they have week-long unplanned downtimes, so using them is okay only as an optional resource.

Even so, there are lots of providers that are cheaper than Amazon and won't screw you over.

Re:Show HN: We moved from AWS to Hetzner, saved 90%, kept IS... by winrid on Saturday June 21, 2025 @01:31PM (#44339221)
Don't depend on a single provider. Always do live replication across providers. OVH downtime is only an issue if your entire company is running on a single server. Split it across a couple zones if this is actually paying bills.
Re:Show HN: We moved from AWS to Hetzner, saved 90%, kept IS... by OutOfHere on Saturday June 21, 2025 @02:10PM (#44339499)
At the very least I backup data to a different provider, so I can restore the services elsewhere, although there may still exist some residual data loss since the last backup. I guess it depends on just how well it's paying the bills.
Re:Show HN: We moved from AWS to Hetzner, saved 90%, kept IS... by sksjvsla on Sunday June 22, 2025 @12:20PM (#44348139)
I have not experienced this in spite of rumours online. As I mention in these two comments, given these we decided to design our way around it by assuming that they would both go down at some point of time (but not at the same time).

1. https://news.ycombinator.com/item?id=44335920#44339234

2. https://news.ycombinator.com/item?id=44335920#44337619