Explore the challenges of running large-scale Kubernetes clusters in this 41-minute conference talk from LISA19. Dive into complex runtime and networking issues encountered when operating Kubernetes in production at scale. Learn how to adapt and tune architectures and component configurations for improved scalability and performance on both control and data planes. Examine ecosystem tools within the context of very large clusters, and discover insights from Datadog's experience managing clusters of up to 2000 nodes. Gain valuable lessons on building mutually beneficial relationships with the Kubernetes community through feedback and upstream contributions.
Overview
Syllabus
LISA19 - Kubernetes the Very Hard Way
Taught by
USENIX