More Power, Less Pain: Building an Internal Platform with CNCF Tools
summary
A production internal platform on Kubernetes for a nine‑person org with a three‑person ops team, built and running with concrete choices: cert manager for TLS, ambassador for ingress, linkerd for the mesh, and argo for CD and rollouts; dev flow relies on telepresence and secrets land in doppler. Zonal clusters avoid the disk‑attach footgun. Observability goes to a vendor offering 1 TB free and the cluster consumes 960 GB of ingest monthly, so vendors handle observability databases CI and certain stateful services.
AI-written summary · about
chapters (17)
- 0:00Introduction
- 1:03UpChieve mission and Kubernetes
- 2:28Building internal platform
- 3:52Cloud-native platform elements
- 6:21CNCF tools vs vendors
- 8:07Kubernetes cluster setup
- 10:35Certificate management with cert-manager
- 11:16Ingress controllers: Ambassador vs Nginx
- 12:17Service mesh: Linkerd vs Istio
- 12:59Argo CD and Rollouts deployment
- 14:01Developer experience with Telepresence
- 15:26Open-source observability challenges
- 16:52Vendor observability and databases
- 18:59Secrets management and CI/CD trade-offs
- 20:46Infrastructure as code dependencies
- 21:50Vendors vs open source
- 22:10Conclusion and key takeaways
related talks
15mMore Power, Less Pain: Building an Internal Platform with CNCF Tools
22mAnatomy of a production grade Kubernetes cluster
26mTwo sides of the platform coin with operator's and customer's voices
27mBuilding a Cutting-Edge Kubernetes Internal Developer Platform at NVID...
41mArchitecting Kubernetes-Based Internal Developer Platforms: Essential Patterns...
28mOvercoming the Challenges of Centralizing Container and Kubernetes Operations
37mBuilding a Managed Platform While Maintaining a Good Developer Experience
31mThe 20 Year Platform – bringing together Kubernetes, 12-Factor and Functions