How we evolved our cloud infrastructure
summary
HashiCorp’s cloud platform moves from a monolithic repo that managed tens of thousands of data planes to a decomposed architecture with componentized repos for network, ingress, cluster and config layers. Automation replaces manual upgrade validation so rollouts drop from weeks to hours, using ephemeral end-to-end testing, metadata tracking and CI-driven deployment across four regions (Western Pioneer + EU expansion). The effort hits major goals in 10 months of a 12‑month mandate and runs monthly failover tests to validate resilience. automation across regions
AI-written summary · about
chapters (22)
- 0:00Introduction
- 1:03HashiCorp and speaker role
- 2:30Original cloud limits
- 5:40Scaling challenges
- 7:25New approach needed
- 8:30Modularizing infrastructure
- 9:33Multi-region automation
- 10:36Automation workflows
- 13:24Innovation after modernization
- 14:30Why 10-month success
- 16:19Deadlines power delivery
- 18:01Executive buy-in importance
- 20:48Self-service future plans
- 21:50Q&A
- 22:12Deadlines vs well-being
- 23:56Multi-region failover
- 24:39Encouraging creativity
- 26:02Missed deadlines handling
- 27:04Multi-cloud strategy
- 27:46Lessons learned
- 29:08Book recommendations
- 30:35Closing
related talks
15mJourney to Hybrid Multi-Cloud
24mHow to survive Cloud – An Ops perspective
27mAdvanced Multipurpose Infrastructure: A Self-Managed Cloud Environment with HashiCorp Terraform
01h 03mContinuous Security in the Cloud
49mAutomation in Deployment on Hybrid Hosting and Private Cloud Environments
38mBuilding a Cloud Native Culture in an Enterprise - Deep Kapadia & Tony Li
25mFrom Heroku to Cloud Native in 7 Weeks
36mPractical Cloud Native: What Works, What Doesn't