Hacker News Reader

They put in great engineering effort to ensure high availability despite random instance failures. I really like their Chaos Monkey.

See: http://techblog.netflix.com/2010/12/5-lessons-weve-learned-u...

One could argue this is how all systems should be designed anyway.

It's cool that Netflix pulled that off, but it seems like a company (like Dotcloud) that could prevent you from building infrastructure on top of infrastructure would be a boon.

As a high-profile EC2 client, I imagine Amazon gives Netflix tons of "complementary" support. A Netflix failure on EC2 would be a big black eye to Amazon PR.

the outages@outages.org mailing list says that Netflix was down today. I don't know if it's ec2 related or not... I was actually hoping someone might have details.

https://puck.nether.net/pipermail/outages/2011-March/002786....