Bpftrace Meets Pixie: Dynamic Monitoring of K8s Clusters Unleashed
Talkat Cloud Native eBPF Day Europe 2022 · on Monday, 2022-05-16· at 11:55· Pavilion 4 | Room A | Level 2 | Central Forum·
summary
Combines bpftrace and Pixie so bpftrace scripts are compiled and deployed across all nodes of a Kubernetes cluster, with printf output intercepted and captured as structured Pixie tables for real‑time queries and dashboards. Demonstrations show running community one‑liner probes (exec snoop, tcp drops), dynamic redeploys across the cluster in seconds, and a visualization that can identify TCP drops visually and surface sudo/exec activity; prototype works with kprobes and tracepoints.
AI-written summary · about
chapters (16)
- 0:00Intro
- 0:25Realtime debugging of Kubernetes
- 1:49Bpftrace
- 2:46Community Support
- 3:43Pixie Overview
- 6:36Relic Architecture
- 7:24Pixie Demo
- 8:22Integration
- 8:32Bpftrace vs Pixie
- 9:31How do we do it
- 10:44Example
- 11:43Community Script
- 12:34Aggregation Scripts
- 13:14Aggregation Script Example
- 16:06Demo
- 24:04Conclusion
related talks
26mTriaging Real-time Security Threats with eBPF-powered Observability
04mInstant, Auto-telemetry for K8s Observability
17meBPF in Microservices Observability
14mFlexible, Open and Easy Observability for Developers
34mTotal Visibility: Harnessing Pixie for a Full Observability Landscape
30mData Science for Infrastructure: Observe, Understand, Automate
27mLow-Overhead Tracing Using eBPF for Observability into Kubernetes Apps and Services
30mObservability in Kubernetes: Unlocking the Power of eBPF