Bpftrace Meets Pixie - Dynamic Monitoring of Kubernetes Clusters

Explore dynamic monitoring of Kubernetes clusters through the integration of bpftrace and Pixie in this conference talk. Learn how to leverage these open-source tools for real-time troubleshooting and performance investigation of applications on Linux systems and Kubernetes. Discover the process of deploying bpftrace programs across cluster nodes, importing existing scripts, and utilizing Pixie's query language for debugging. Witness live demonstrations showcasing the identification of TCP issues and detection of unauthorized bitcoin mining in K8s clusters. Gain insights into Pixie's architecture, community support, and the advantages of combining bpftrace with Pixie's monitoring approach. Understand the differences between bpftrace and Pixie, explore aggregation scripts, and see practical examples of their implementation.

Intro
Realtime debugging of Kubernetes
Bpftrace
Community Support
Pixie Overview
Relic Architecture
Pixie Demo
Integration
Bpftrace vs Pixie
How do we do it
Example
Community Script
Aggregation Scripts
Aggregation Script Example
Demo
Conclusion