Explore performance insights into eBPF in this 35-minute conference talk by Dmitrii Dolgov from Red Hat. Gain a step-by-step understanding of how to analyze and optimize eBPF programs, including collecting execution metrics, profiling techniques, and avoiding common pitfalls. Learn about adding extra instrumentation probes, improving performance through BPF instruction sets, implementing BPF to BPF calls and tail calls, and utilizing batch operations. Discover advanced concepts such as park allocators, Bloom filter maps, and task log storage to enhance your eBPF implementation skills and achieve better system visibility.
Performance Insights into eBPF Step by Step
Overview
Syllabus
Introduction
Background information
What are we doing
Adding extra instrumentation probes
Improving performance
BPF instruction sets
How to pin instruction sets
BBF to BPF calls
BPF to tail calls
Batch Operations
Park Allocators
Bloom Filter Map
Task Log Storage
Conclusion
Taught by
Linux Foundation
