Environments as Cattle
summary
Advocates treating entire environments as disposable cattle that are fully modeled and versioned so each feature or customer gets its own temporary instance to validate deployment, configuration and schema migrations. Uses Terraform/CloudFormation/ARM/Helm plus Ansible/Chef/Salt, Docker, Jenkins and DB tools like Flyway/Liquibase/Alembic to automate build, snapshot data, test integrations and tear down. Claims eliminate frequent non‑prod outages (example: client cut a two‑engineer, eight‑week environment build to a 30‑minute Jenkins job, a 1280× reduction) and recover 40–104 lost engineer hours caused by ~8 hours/week non‑prod downtime.
AI-written summary · about
chapters (19)
- 0:00Introduction to talk
- 0:24Pets versus cattle analogy
- 1:48Speaker background
- 3:14Cost of non-prod downtime
- 5:23Effects of unplanned work
- 7:17Traditional CI/CD pipeline flaws
- 9:45Shifting left and testing deployments
- 12:34Getting started with versioning
- 14:22Infrastructure management
- 15:31Configuration management
- 16:39Database management
- 18:30Automating default data load
- 19:53Deployable release components
- 21:59Developer environment benefits
- 22:40Customer environment advantages
- 23:47Testing deployments not just code
- 25:11Environment rebuild efficiency
- 26:14Tips for implementation
- 29:37Q&A session
related talks
27mAdvanced Multipurpose Infrastructure: A Self-Managed Cloud Environment with HashiCorp Terraform
01h 01mTaming the Modern Data Center
44mTaming the cloud sprawl
27mCase Study: Re-Thinking Our Infrastructure Tooling
05mAn Environment Per Branch with Infrastructure as Code
27mApplication-Driven Infrastructure
48mMoving from spaghetti infrastructure to composable cloud environments
43mEvolutionary Infrastructure Using Pipelines