Hacker News Reader

Comment by ollien | Hacker News Reader

Hug of death? Archives don't have it either

It's demoralizing that we live in 2026 and a modest level of traffic is still enough to take down half the web.

Google shutting down it's public cache made it even worse.

I've worked with a lot of "smart" engineers over my career with whom autoscaling is a completely foreign concept, or a half-baked concept at best. My favorite is devs pushing to migrate to k8s because it "scales" and won't hear a single thing otherwise.

You don't need autoscaling for a basic blog to handle the traffic of Hacker News. People have had "this page is hosted on <vintage computer>" sites hit the front page that have stayed up.

why do you need autoscaling for an (essentially) static website?

For a static blog you dont need that. My blog made the (near bottom of) HN front page last year. I didn't notice a thing until I looked at the traffic report (using goaccess). Static site generator with the result served from a bottom of the barrel Hetzner VPS, using nginx.

If a blog needs "scaling" of any sort you are doing it wrong.

Even hacker news itself runs on a single server without any scaling.

Look at the techempower benchmarks, even python stuff can handle 40k requests per second involving database lookups, and when things are cachable compiled/JITed languages scale to millions on 2 CPUs. And if you object that those are microbenchmarks, I have seen this in real systems too. E.g. converting a convoluted hairball of python, lambdas and SQS to a Rust service running on 2vCPUs reduced latencies from seconds to milliseconds.

The "If necessary I can add autoscaling later" mindset is what leads to bloated, slow websites.

Agree.

I don't think folks think clearly about how their setup scales computationally when they are choosing their stack. I've been writing a lot of software designed for drop-dead simple self-hosting lately (can run CGI-bin or as a standalone server, based on the env it's invoked in), and I put my own instances behind CGI, keenly aware of the risk on the open internet. But I design .htaccess so that the main server is hitting caches for anything anonymous. This works Just Fine for hitting the HN front page in my experience on a 1GB VPS, since the request never even makes it to my Python code.

I think folks just don't prioritize these sorts of optimizations.

Looks like the IP address is from Linode. So it could be a $5 VPS running an inefficient server-side framework.

I currently host my personal Web site as static files via AWS CloudFront. But even static hosting on a $5 VPS with Nginx might handle an HN surge fine.

I host my AI-generated code portfolio site separately (to unambiguously distinguish it from my 100% human-written open source work), using github.io (seemed appropriate) with a custom domain.

>static hosting on a $5 VPS with Nginx might handle an HN surge

My personal site is hosted on this exact setup. Twice I made the homepage, twice I noticed no slowdown whatsoever.

The only problem is that in May of this year, when my SimCity 3k in 4k article was found by the HN crowd, its access.log bloated to 200Mb causing my logrotate script (that hooks in Goaccess) to blow up every time since when trying to process it, and this log has been growing every day since - because logrotate isn't doing much rotating!

When I can massage my VPS into processing it, I'll have the stats to enjoy.

The resource has died for me as well, but I used to read it when the post has appeared.

Reply on news.ycombinator.com