5 ms·

Some stats from my front-page a month ago: - Time on front page: 15 min - Views in first 15 min: ~5000 - Views in first 60 min: ~7000 - Views in firs

Some stats from my front-page a month ago: - Time on front page: 15 min - Views in first 15 min: ~5000 - Views in first 60 min: ~7000 - Views in first 600 min: ~11,000 - Views over the next 5 days: ~5000 HN front pages bring a bunch of users quickly, but it isn't a scale that will cripple moderate hardware with some basic performance tuning (directly serving assets, for example). The tail is also great, fully 3/4 of my post-front-page views come from click-throughs on social.

Nextgrid 8y ago
It actually seems pretty manageable, I'm surprised the site died. You're running Wordpress by any chance?
- misframer 8y ago
  I’m using Hugo and Caddy. It’s running out of file handles.
  - mholt 8y ago
    The hugo plugin, or are you running hugo separately?
    - misframer 8y ago
      Separately. I haven't updated it to use the plugin. Thanks for Caddy, BTW!
      - mholt 8y ago
        File a bug, yours is the first we've heard of something like that. And you're welcome!
majewsky 8y ago
I saw similar numbers on my server. About 10,000 views over the course of an entire day (derived as total data transmitted divided by the size of one full page load), with the bulk concentrated in the first two hours. My nginx served it without breaking a sweat. Peak CPU usage was at 5% and network load peaked at 1 Mb/s (data at 60-second resolution). My stack was just nginx serving static files, on a 1/1 VM that also hosts half a dozen other services (XMPP, Gitea, Matrix, Mumble etc.). It is my belief that everyone who encounters the HN hug of death is doing something wrong.
- Macha 8y ago
  I used to have a blog that ran on Wordpress (3.0 was I think the current version at the time, this was a while ago) with the APC plugin, served via Apache. An article hit HN and it fell over. Some time later I had migrated to blogofile and nginx and made it to HN again, the server barely noticed.

Nextgrid 8y ago

It actually seems pretty manageable, I'm surprised the site died. You're running Wordpress by any chance?

misframer 8y ago
I’m using Hugo and Caddy. It’s running out of file handles.
- mholt 8y ago
  The hugo plugin, or are you running hugo separately?
  - misframer 8y ago
    Separately. I haven't updated it to use the plugin. Thanks for Caddy, BTW!
    - mholt 8y ago
      File a bug, yours is the first we've heard of something like that. And you're welcome!

majewsky 8y ago

I saw similar numbers on my server. About 10,000 views over the course of an entire day (derived as total data transmitted divided by the size of one full page load), with the bulk concentrated in the first two hours. My nginx served it without breaking a sweat. Peak CPU usage was at 5% and network load peaked at 1 Mb/s (data at 60-second resolution). My stack was just nginx serving static files, on a 1/1 VM that also hosts half a dozen other services (XMPP, Gitea, Matrix, Mumble etc.). It is my belief that everyone who encounters the HN hug of death is doing something wrong.

Macha 8y ago
I used to have a blog that ran on Wordpress (3.0 was I think the current version at the time, this was a while ago) with the APC plugin, served via Apache. An article hit HN and it fell over. Some time later I had migrated to blogofile and nginx and made it to HN again, the server barely noticed.