Netflix is so great when it comes to sharing how they do what they do techwise. I expect this post will be evolving over time.
Key takeaway: Innovation and reliability before efficiency (cost or performance)
This doesn’t mean efficiency is ignored. In fact they have different strategies to encourage that on its own. Resource pools and sharing, heavy investment in long term reservations, reports that go to dev team managers on how their teams are doing in terms of cost and alarms when provisioned capacity changes in a big way.
Really awesome stuff.
- https://www.slideshare.net/cpwatson Coburn Watson, Director Cloud Performance and Reliability @ Netflix
- Goto; Keeping your cloud footprint in check