Linux performance: tools and methodology
Summary
Brendan Gregg brings together performance models, checklists, diagrams and Linux tools for systematic diagnosis instead of random sequences of commands. The overview links to numerous free talks and tools. USE stands for utilisation, saturation and errors.
Ideas
- USE checks every resource for utilisation, saturation and errors.
- Flame graphs condense samples into readable call paths.
- perf observes CPU, kernel and applications with shared tools.
- BPF allows safe, programmable measurements on a running system.
Insights
- A fixed methodology reduces confirmation bias during troubleshooting.
- Performance problems often arise at the boundaries between system layers.
- Visualisation reveals distributions that averages hide.
Facts
- Brendan Gregg developed the flame graph visualisation.
Recommendations
- Save measurements before every change to the system.
- Start with the resources and then follow the proven bottleneck.
References
Links to the original source and the Web Archive open in a new tab.