DevOps in the Cloud
summary
Case studies show Amazon teams run a build it run it model where many independent two‑pizza teams own services deployed with infrastructure as code and reviewed blueprints. Examples: the detail page makes hundreds of parallel widget calls so pages gracefully degrade; search uses Fault Injection Service (FIS) plus an orchestrator, SLO guardrails and an andon emergency lever to validate resilience under load (search hit ~84,000 rps on Prime Day across 40+ backend services); relay goes multi‑region with Lambda and DynamoDB Global Tables.
AI-written summary · about
chapters (13)
- 0:00Introduction and background
- 1:16Breaking the devops wall
- 3:03Two Pizza team model
- 5:29Undifferentiated heavy lifting
- 7:15Amazon’s service architecture
- 7:58Amazon detail page example
- 9:40Amazon search and chaos engineering
- 13:10Building on Fault Injection Service
- 18:00Amazon Relay app architecture
- 21:12Multi-region resilience
- 23:40Builder tools and automation
- 24:44Infrastructure as code
- 25:46Summary and key takeaways
related talks
31mDevOps in the Cloud: Case Studies of Amazon.com teams and their resilient architectures
21mDevOps Culture at Amazon
28mContinuous everything: Moving to DevOps on AWS
48mDevops@Scale
23mCI/CD at Scale: Best practices with AWS DevOps Services
15mDevOps at scale
16mDevOps challenges in a microservices world
01h 02mTransforming IT pros to DevOps gurus: Ssecure your new tech stacks