Infrastructure Automation (anti) patterns
summary
Claims that infrastructure automation succeeds when it targets reproducibility and ownership, and fails when it’s scattered in home directories, UIs or long‑running branches. Recommends provisioning with Terraform, desired‑state enforcement with Chef/Puppet and orchestration with Ansible/Salt, running configs regularly (every 20 minutes) to prevent drift, and storing only user data for backups. Warns against golden images, warns that automating only the OS leaves app state unmanaged, and advocates package-config-service patterns and a single source of truth.
AI-written summary · about
chapters (8)
- 0:00Introduction and background
- 2:08From manual to reproducible infrastructure
- 7:22Infrastructure as living code
- 13:16Common automation patterns
- 21:02Stacks and environments best practices
- 25:55Anti-patterns in provisioning
- 29:07DSL vs full language trade-offs
- 30:31Lessons learned and Q&A
related talks
01h 02mTools and Techniques for Infrastructure as Code
27mApplication-Driven Infrastructure
16mDeclarative Infrastructure Basics for Systems Admins
33mScaling Infrastructure as Code
13mNavigating the infrastructure skills gap for DevOps teams
35mHow to Make Your Automation a Better Team Player
01h 02mTools and Techniques for Infrastructure as Code
15mInfrastructure as a Code: Best Practices and Myths