Overview
Explore Berkeley Packet Filter (BPF) applications for chaos engineering and tracing in Kubernetes environments. Delve into the world of distributed computing and learn how to analyze and simulate kernel failures using BPF, a flexible and efficient virtual machine-like construct in the Linux kernel. Discover practical solutions for handling unpredictable faults and performance issues in cloud-based clusters. Gain insights into developing BPF programs to analyze kernel-caused performance problems and conduct kernel chaos experiments on Kubernetes. Learn about the process of converting BCC-based BPF applications to achieve "Compile Once – Run Everywhere" functionality on 5.x kernels. Understand the importance of BPF in addressing complex issues in distributed systems, with a focus on its application in TiKV, a distributed and transactional key-value database.
Syllabus
BPF for Chaos and Tracing in Kubernetes - Wenbo Zhang, PingCAP
Taught by
CNCF [Cloud Native Computing Foundation]