The HTML source is pretty good too. A doctype that's half HTML5, half XHTML. A pre-IE7 script tag thing. A meta-keywords tag. Lists that don't actually use any list tags. Raw PHP tags being dumped into the HTML output. No closing body or html tags. This joke has _layers_, man.
> Don't like uploading your files via FTP? No problem! Send us your files on a floppy and we'll upload it for you. It's one of the many ways we do our best to accommodate your needs. Just remember, our floppy system does not accept Windows, Linux, or Mac floppies.
I like how deep the rabbit hole goes on pauperhosting. The accidentally included admin panel with leaked customer db, the VPS demo, everything. I'm pretty sure I haven't even discovered half of it yet.
Almost all the customer names in that excel file are taken from "huilende rappers - waddepjedangedaan", on a joke rap album by drum and bass legends noisia. It's all in Dutch... but the made up names are so incredibly funny that i use them for dummy data everywhere.
This looks like a good service for compute, but it's way over priced for storage. It probably makes most sense to use the included disk for application code only and store all data in something like S4 [1]
How bad is the HN hug of death btw? My laptop can serve 200k QPS per core. I can't imagine HN is that much more intensive? Are 1vCPU VMs just that much worse?
Yes, they are much worse than bon-fide actual single cores on most recently purchased laptops. But also, the HN hug of death mostly happens with tiny non-static sites that are sloppily hosted. If you're serving static assets it's hard to "hug of death" a Casio watch.
M3 at work, Acer Swift 3 at home. Both are comparable in that regard. You can do 200k QPS of actual work (a little protobuf parsing, a little old-school ML, handling the networking, a little HTTP1.1 parsing, ...), more if you just want vanity metrics, just by wrapping something like uSockets [0] and not doing anything to explicitly pessimize the system.
You can do better with a hand-crafted solution, but most projects don't need anything fancier.
How do we know they even power up any servers at all, for the service? I mean, they could simply just not power up any servers at all. Till then, I’ll opt for the ‘take-my-money’ plan.
Ideally they wouldn’t need to power up the servers but you can see from the graph that unfortunately there’s still some uptime, though they are working on it, maybe once they reach 100% downtime
Out of interest, that mentions "Dedicated Win 3.11", so silly question, I assume you would be able to install DOS on a dedicated server, but might it be possible to run DOS/Win 3.11 on EC2 (I couldn't find a publicly accessible list of AMIs)
