The Evolution of GitOps in Platform Engineering
summary
Charts the shift from Git as a human collaboration layer to limits at scale: file-based, pull-driven reconciliation creates config sprawl and breaks when thousands of agents hit git and provider APIs. Shows concrete pivots—OCI-based delivery and “gitless GitOps” that push immutable artifacts—and cites scale tests (managing 15,000+ clusters, deploying ~5,000 apps or 1,000 clusters in under two hours with a ~€200k budget). Argues tools like Argo CD, Helm and Kustomize enable scale but hide real cluster state, so platforms need an operable state store, supply-chain visibility for the Cyber Resilience Act, and final manifests for AI ops.
AI-written summary · about
chapters (16)
- 0:00Introduction
- 2:27Origins and early challenges
- 4:15Agent-based GitOps principles
- 7:07Scaling with fleet management
- 9:13Git vs OCI as intent source
- 11:25Multi-cluster management architectures
- 13:55Security and pull-based benefits
- 14:35Platform engineering integration
- 17:01Adoption challenges and costs
- 19:28Cultural shifts and education
- 21:15Progressive delivery direction
- 22:42Resilient and disposable platforms
- 23:44GitOps engines comparison
- 25:30AI agents in platform engineering
- 27:16AI challenges and regulations
- 29:00Q&A
related talks
30mThe Objectives And The Principles Behind GitOps
40mHow GitOps Changes Everything in CI/CD
26mKeynote Session: GitOps as an Evolution of Kubernetes
02h 47mDes silos au Platform Engineering en passant par le DevOps : adopter GitOps et aller au-d...
28mManaging GitOps as a Platform Service
55mGitOps, Fostering Next-Generation Operations Teams
09mThe Future of GitOps in Modern DevOps
29mPutting the Ops in GitOps - New Challenges in Managing Cloud Native Environments