From Silos to DevOps to Platform Engineering: embracing GitOps and going behind the hype
summary
Argues that Platform Engineering pairs GitOps single source of truth with dedicated platform teams and operators to cut developer cognitive load, not just reshuffle responsibilities. Describes building an internal developer platform on Kubernetes that uses operators to manage external databases, Argo CD for declarative delivery, and a “golden path” developer UX; early rollouts yield faster onboarding and fewer ad‑hoc infra choices but warn that a poorly run platform becomes another layer of complexity.
AI-written summary · about
chapters (18)
- 0:00Platform engineering journey
- 4:20IT complexity over 30 years
- 10:27Tooling and dev responsibilities
- 12:39DevOps anti-patterns
- 15:56Cloud adoption consequences
- 20:12Containers and microservices
- 22:42Kubernetes intern you never wanted
- 25:59Cognitive load of cloud services
- 27:50What platform engineering solves
- 30:41Three pillars of internal platforms
- 31:24GitOps as single source of truth
- 33:55Platform teams with Team Topologies
- 37:04Golden path for developers
- 38:52Platform as a product
- 41:02Value of glue in platform engineering
- 43:52Two platform engineering examples
- 47:28PaaS without Kubernetes
- 50:02Closing remarks and Q&A
related talks
47mGitOps - Using Git as your source of truth for build, deploy and observability
34mGitOps for the people
44mIs Platform Engineering a hype? Let's build a platform together!
14mOvercoming Kubernetes Complexity with Platform Engineering
45mGitops: The Kubernetes way
21mGitOps Dynamics: Navigating the new era of DevOps
31mGitOps and Kubernetes: a Winning Combo
35mGitOps - the holy grail of devops