before this issue, I hadn't given it much thought either. Most of the stuff I run on nixnet.services
is on the same VPS from netcup. Their marketing pages implied *to me* that, on average, a VPS would have about gigabit connectivity. A month or so ago, we noticed that all services were incredibly slow and having odd issues. I assumed it was just some application being hit particularly hard so I planned to implement more aggressive caching and reduce that a bit. A few days later, issues still present, decide to run bmon on the off-chance it's something related to bandwidth, and find out that the server has less than 100 MiB/s connection. Contacted support and found out that another customer on the same host seems to be using all the available bandwidth and there's nothing they can (will) do about it.
So I don't actually know how much bandwidth we need, just that 100 MiB/s is far too little. We're in the process of moving everything into LXD and onto the AX51-NVME from Hetzner. Pleroma should be the next one to go this weekend.