Airgapped Kubernetes - Hannes Probst
summary
Demonstrates that production Kubernetes can run isolated by routing all package and OCI pulls through a central artifact store (JFrog/Nexus/Harbor) and using containerd mirror configuration plus cloud-init from a templating engine (osm) to provision nodes. Shows a working air-gapped multicluster on AWS reference with Cluster API machine provisioning, recommends avoiding heavy pre-baking, prefers flexible boot-time logic, and describes offline replication options (registry tar via USB). Claims reduced attack surface and predictable performance and the reference implementation works.
AI-written summary · about
chapters (7)
- 0:00Introduction to airgapped Kubernetes
- 2:32Reasons for airgapped Kubernetes
- 5:46How to implement airgapping
- 10:46Provisioning and templating engines
- 15:26Central artifact store architecture
- 17:40Key takeaways and benefits
- 18:45Q&A session
related talks
18mManaging Kubernetes in Air-Gapped Environments
35mManaging Kubernetes in Air Gap/Offline Environments
35mManaging Kubernetes in Air Gap/Offline Environments
45mNo Internet - No Problem? Air-Gapped Kubernetes on Bare Metal
33mOffline but Not Blind: Observability in Air-Gapped Kubernetes Environment
25mCloud Native Deployments in Air Gapped Environments
36mThe Architecture of a Multi-Cloud Environment with Kubernetes
34mKubernetes in Highly Restrictive Environments