High Performance Tracing in Production

Explore high-performance tracing techniques for production environments in this 41-minute Linux Foundation webinar. Dive into the underutilized suite of Linux dynamic tracing tools to analyze and understand system bottlenecks across diverse workloads and technology stacks. Investigate common performance problems through demonstrations, flame graphs, and various tracing tools. Learn about sample-based profiling, dynamic instrumentation, safe facilities, and eBPF. Discover the Linux performance subsystem, continuous profiling in production, and performance indicators. Gain insights on cross-server fleets, kernel probes, and data cleansing. Enhance your ability to optimize system performance in complex, modern environments with multiple services and technologies.

Introduction
Sample Performance Problem
Demo
Recap
Flame Graphs
Common Performance Problems
High Level View
Tracing Tools
Performance Mapping
Sample Based Profiling
Perf
Dynamic Instrumentation
Safe Facilities
EBPF
Asynchronous Communication
Summary
Linux Performance Subsystem
Continuous profiling in production
Does running the profiler add performance overhead
Performance indicators
K Probes
Best Way to Trace RT
Apache Bench
Books
Monitoring tools
Data cleansing
Crossserver fleets
Kernel EPPF