How Big Is Too Big? Rethinking Kubernetes Cluster Scale

Explore the practical and architectural limits of Kubernetes cluster scaling in this 15-minute conference talk that helps teams understand when to grow a cluster and when to split it. Learn that while Kubernetes can support clusters with thousands of nodes, scaling that far isn't always advisable due to increasing operational complexities including slower upgrades, API server strain, expanded blast radius, and multi-tenancy challenges. Discover a strategic framework for thinking about Kubernetes cluster scale that examines key trade-offs such as control plane performance, failure domains, and upgrade complexity. Address the challenges of securely supporting multiple tenants and evaluate when multi-cluster strategies make sense. Focus on pragmatic approaches to scalability rather than theoretical limits, gaining insights to design clusters that are not only large but also resilient, efficient, and maintainable for long-term operational success.