Infrastructure and AWS at scale: The story of Posti
summary
Claims a reproducible multi-account Landing Zone that lets 20+ teams run ~200 production services for 4+ million users with a small infra staff, and shows the actual repo layout, CI flow and billing mapping used. It uses AWS Organizations, CloudFormation-based Landing Zone 2.0 patterns automated by Skepta and GitHub Actions, fronted by a custom Upsync gateway and moving heavy event traffic to Kafka. The result: per-service infra repos, one infra team owning templates and teams owning deployments, which materially reduces support overhead and scales.
AI-written summary · about
chapters (14)
- 0:00Posti infrastructure challenges
- 3:35Posti's scale and structure
- 8:10Problems with 50 teams and 200+ services
- 10:59Why simple diagrams fail
- 13:26Multi-team AWS complexity
- 16:35Cost tracking and accountability
- 19:28AWS Landing Zone introduction
- 23:00AWS Organizations to Landing Zone 1.0
- 25:29Landing Zone 2.0 improvements
- 27:13Skepta framework choice
- 29:20Skepta automates AWS deployment
- 32:09Team Landing Zones structure
- 35:18Team responsibilities and support
- 37:03Security in multi-account AWS
related talks
45mThe Automation of a Landing Zone
23mBuilding Self-Service Development Platforms
28mBuilding K8s on AWS & Enabling Developers to Deliver. What to Do Better - Greg Sansom & Fahad Tariq
43mReplatform to Amazon EKS
34mInfrastructure at Scale Using ArgoCD
33mSpace Invaders for Kubernetes: Consolidating 50 Clusters at Scale
35mGitOps in Action
30mFrom API Events To Production in Minutes: Backstage as the Front Door...