Frequent storage breaks, anyone using this for production?

[carstenblt]

Since I started using hetzner with this project, I don't think there have been more than 2 months without a break that needed manual interaction.

• up until recently, hetzner had random problems from time to time leading to longhorn breaking. It was always just one or two nodes, and cordoning them helped. It was sad to see that longhorn was not able to cope with this - it just endlessly tried to mount volumes on broken nodes. A few months ago they had a few days when almost every node broke and they seemed to have fixed the problem all together after that.
• a couple of months ago MicroOS broke the NFS package. Manual rollback was required.
• two weeks ago MicroOS broke is said, manual rollback required.
• this week a rollback was needed again, I didn't check the reason, but the issue was resolved.

I am about to give up. Maybe disabling auto-updates and applying them manually from time to time might work.

Is anybody actually using this for a production setup?

I'd be more than willing to pay like $50 a month just for a stable release channel where all updates are thoroughly tested (if hetzner actually fixed their problems)