DevOps Culture at Amazon
summary
Amazon describes a concrete DevOps practice: you build it you run it with small two‑pizza teams, microservices each owning their data and APIs, and multiple automated pipelines per service (app, infra, OS). Pipelines enforce code review, unit and integration tests, gamma canary and one‑box rollouts. Resulting scale is specific: 190 million deployments in 2019, CloudWatch handles about one quadrillion metric observations per month, and Systems Manager heartbeats over 10 million instances.
AI-written summary · about
chapters (12)
- 0:00DevOps culture at Amazon
- 0:46Amazon innovation flywheel
- 3:57Enterprise IT vs Amazon model
- 6:22Monoliths to microservices
- 8:12Microservice architecture best practices
- 9:34Teams and Conway’s law
- 11:21Two-pizza teams and pipelines
- 12:24DevOps pipeline benefits
- 14:08Code review and deployment
- 18:01Automated pipeline stages
- 18:22AWS management services
- 20:05Team ownership and innovation
related talks
27mDevOps in the Cloud
43mThe Amazon Software Development Process
31mDevOps in the Cloud: Case Studies of Amazon.com teams and their resilient architectures
28mContinuous everything: Moving to DevOps on AWS
23mCI/CD at Scale: Best practices with AWS DevOps Services
47mWhy We Need DevOps Now: A Fourteen Year Study Of High Performing IT Organizations
23mCI/CD at Scale: Best Practices with AWS DevOps Services
28mDevOps in a Data Warehouse: Inside Out