How Netflix Thinks About DevOps
summary
Netflix rejects traditional DevOps guardrails and opts for a culture-first model that gives engineers full production access and expects freedom and responsibility. Systems run on AWS with Akamai and Netflix Open Connect; deployments use Spinnaker with canary rollouts (10% first). Scale: ~200M members, 190 countries, thousands of microservices, tens of thousands of instances, and 2.5 billion metrics per quarter. A concrete build is Netflix Drive, a C++-backed studio filesystem that shards large files to sync only changed chunks.
AI-written summary · about
chapters (14)
- 0:00Introduction and welcome
- 1:43Speaker introduction and agenda
- 2:24Netflix’s scale and architecture
- 5:54Netflix’s approach to DevOps
- 10:49You build it you run it principle
- 13:14Culture drives DevOps
- 14:39Q&A begins
- 15:00Tooling and documentation
- 17:14DevOps for new grads
- 19:01Monitoring and observability
- 20:47Cloud knowledge for engineers
- 22:09High-level service design
- 24:39Language flexibility at Netflix
- 25:40Final thoughts and closing
related talks
31mHow Netflix Thinks About DevOps
26mBeyond the Culture Deck What You Don’t Already Know About Netflix
27mBehind the Scenes: Running Critical Services @ Netflix
58mBuilding a Culture of Continuous Delivery
29mFast, Safe, and Continuous -- Picking all 3 at Netflix
50mFast Delivery
30mA conversation with Netflix
01h 05mSpeed & Scale: How to Get There