Lessons Learned as a DevOps Consultant
summary
Lists concrete practices and failures: run a weekly open‑mic to reuse Terraform/Pulumi modules and template pipelines, and require written acceptance criteria to avoid rework; expect surprises — a Terraform policy rename removes KMS rights and takes a managed Kubernetes control plane offline for ~18 hours; replicate account‑level IAM in a separate test account before changing policies; never change production when fatigued (deleted Oracle redo logs once cost an extra hour); prefer the right architecture (serverless vs Kubernetes) over the popular one.
AI-written summary · about
chapters (9)
- 0:00Introduction
- 2:33Collaboration over silos
- 5:42Cloud infrastructure surprises
- 10:21Avoid single-tool lock-in
- 13:19Never assume requirements
- 15:26Popular tools aren’t always the right fit
- 18:59No production changes tired
- 22:35Embrace fast change
- 27:47Q&A
related talks
22mLessons Learned from Scaling Infrastructure as Code
05mLessons learned from building a cloud-native platform
50mDevOps Project Failures - Lessons Learned and Strategies for Recovery
28mLessons Learned from Writing Thousands of Lines of IaC
35mTerraform Practices - The Good, the Bad and the Ugly
42mInfrastructure as Code - Lessons learned from Dev to Ops
32mTales from the DevOps Transformation Trenches
41mTerraform Practices - The Good, the Bad, and the Ugly