Posted by tradertef 21 hours ago
Unfortunately, this isn’t something that shows up on spec sheets when you’re choosing a service. :-/
You probably won't see this unless both the following are true for your situation:
1) You have a workload that makes this issue noticeable. Long-lived connections and large transfer sizes make it more likely you'll notice. Loading 20kb of static html over the connection likely won't seem to have any problems (unless you run repeated trials and network analysis tools). Of course, modern websites can be pretty large...
2) Your users are long-term enough and in communication with you so these issues can even be noticed in the first place. Also helps if they're technical. If you're not hearing the story and aware of the situation on the other end of the line, all you see is a slow connection, could be anything causing it, and there are plenty of them for reasons that have to do with things closer to the client's end.
So all e.g. an e-commerce site might see is a somewhat higher bounce rate than necessary (due to some fraction of their users experiencing the site like it's on a somewhat-jittery ISDN line) without even knowing they're leaving money on the table because they likely have no way of even being aware of the problem.
[EDIT] Yes, we tried shifting around a bunch of ways on DO's side trying all kinds of ways to fix this, I'm quite sure it wasn't that we were unlucky with our hardware draw there or just one of their datacenters had this problem. It was something past the edge of their network.
Here’s the more honest one i wrote a while back:
> Infrastructure: Bare Server > Containers > Kubernetes
The problem with recommending a bare server first is that bare metal fails. Usually every couple of years a component fails - a PSU, a controller, a drive. Also, a bare metal server is more expensive than VPS.
Paradoxically, a k3s distro with 3 small nodes and a load balancer at Hetzner may cost you less than a bare metal server and will definitely give you much better availability in the long run, albeit with less performance for the same money.
IME, hardware is much more reliable than people think.
You do get Opus 4.6, and it's really affordable. I usually go over my limits, but I'm yet to spend more than 5 USD on the surcharges.
Not seen a reason to switch, but YMMV depending on what you're doing and how you work.
The pricing is so good that it's the only way I do agentic coding now. I've never spent more than $40 in a month on Opus, and I give it large specs to work on. I usually spend $20 or so.