I've emailed Patrick, so let's hope he can bring it back up soon.
It's a shame sites can't handle HN-level traffic automatically. My own site got borked when I put it on here without knowing much about web server configurations. Are there some standard practices which one should know?
Basic VPS (on stable provider) + nginx + php-fpm + eaccelerator, barely noticing any kind of traffic spikes (handled SlashDot, HN, HabraHabr, Reddit, Delicious, many others).
Of course it depends on how intensive your app is, but for moderately loaded site - this combination is more than enough. For ex. I've tested Wordpress with this configuration - it can handle nearly 70 mil. hits per day (including into this number hits to images, so probably around 10 mil. actual visitors).
Wouldn't recommend for real-time upgrade (during spike). For real-time go for this: http://gist.github.com/638751 (if load average > 5 - sorry, exit). Crude, cruel, but helped me survive being unexpectedly TechCrunch'd on Apache (didn't know about nginx at the time), then you can finally edit your files, go for caching of nearly everything.
This is a really neat idea, but it seems to track how fast the mouse moves over the canvas instead of the amount of time the mouse is over parts of the canvas.
hey guys its up again, I didn't know that my webserver wouldn't be able to handle such an amount of traffic. I reconfigured the webserver, now it should be ok.
I'm going to add iPhone/iPad touchmove support for the canvas heatmap example in the next days. cheers, Patrick
