[christophilus]

Well, unrelated to the topic, but I'd be curious how many request / sec you were getting, and how many nodes Render autoscaled up to. I'm running a client's site on a single Render node right now and have been curious how it'd stand up to traffic spikes.

[Andrew_Russell]

I can't tell from the logs that I can see. But the site is all or almost all text from Render with images on AWS, and bandwidth topped out at "220MB" (per hour? Not sure) from Render this morning. It came back up as soon as I enabled scaling, and it only scaled to 3 nodes.

It's also a Python 3 instance running Django/Wagtail/Puput, and all blame goes to me for poor coding. If anyone reading this happens to be a Wagtail & Render expert and wants some short contract work fixing my code, please reach out ([e-mail])!

[gregsadetsky]

I recently had an unoptimized Django site running on render.com with the `plan: standard` setting survive the HN front page with a single node. I had autoscale set to on, but it never needed to scale up.

I don't know the requests/second, but at peak on Google Analytics it said that 300 people were on the site "right now" -- unsure of what that corresponds to. ~5-10 reqs/s roughly?

Also note that render.com puts Cloudflare in "front" of the nodes automatically, which helps with some caching.

(I'm unaffiliated btw, just moved to them from Heroku and have been happy to far)

[judge2020]

> Google Analytics

Hardly effective for gauging HN readership, given the abundant use of ad blockers. Based on no evidence whatsoever, 100x that and you should have a reliable estimate.