This session shares real-world lessons from running Kubernetes far from the data center. I intend to discuss specialized distros and tools: e.g. using lightweight Kubernetes like K3s or MicroK8s, and projects like KubeEdge to extend clusters to edge nodes.
I will cover network challenges (intermittent connectivity, high latency) and how to design clusters (single-master vs multi-master, backup control planes) to stay resilient.
For example, I will highlight techniques that allow an edge cluster to keep operating when disconnected from a central control plane, citing that edge computing improves reliability by letting “critical work… be done” even if links fail. I also touch on remote management: using GitOps and management platforms to automate upgrades and configurations at scale. Attendees will see how to balance consistency and resilience when K8s meets IoT/edge demands.