Optimizing Linux Systems: Leveraging Guider for Advanced Performance Analysis and...

on YouTube ↗40m

" summary

An open source tool named Guider provides automatic performance monitoring as a background daemon that reads config thresholds, runs commanded investigations, and produces text reports, SBG visualizations and flame graphs; it ships as ~153 commands in a ~2 MB package, uses ~2–3% of one core while monitoring, and a 3 megabyte ring buffer yields about 15 minutes of captured system history. Collected JSON and reports aggregate across devices for histograms and comparative analysis and are already used to find and fix real-world automotive platform issues.

AI-written summary · about

" chapters (33)

1. 0:00Performance analysis challenges
2. 1:06Platform performance issues
3. 2:33Performance tools overview
4. 3:37Guider development
5. 5:00Guider installation
6. 6:26Workloads and monitoring
7. 8:56File descriptors and sockets
8. 9:41Function monitoring with utap
9. 11:14System call monitoring
10. 12:00Python function tracing
11. 12:23System profiling with rec
12. 13:49Task scheduling timeline
13. 14:13System call profiling
14. 15:19System call tracing
15. 16:00Advanced function handling
16. 16:44Python tracing with fytrace
17. 17:04Signal tracing
18. 17:26Memory leak detection
19. 18:30Automotive constraints
20. 19:14Guider as daemon
21. 21:18Monitoring thresholds
22. 23:48Performance reports
23. 25:36Snapshot details
24. 27:02Report visualization
25. 28:48Reports to flame graphs
26. 29:30Guider shell control
27. 30:13Multi-device data
28. 31:38Remote data transfer
29. 32:20Trend visualization
30. 33:46Comparative analysis
31. 35:11Analyzing anomalies in large performance datasets
32. 37:21Guider future plans
33. 38:21Q&A

" related talks

• 53mGuider: Performance Tracing Features for LinuxPeace Lee · Open Source Summit + Embedded Linux Conference + OSPOCon 2021
• 51mLinux Systems Performance - ArchitectureBrendan Gregg · Percona Live 2016
• 41mLinux Systems PerformanceBrendan Gregg · LISA19
• 20mFrom Debugging to Prevention: Smarter Linux Diagnostics with AutomationAkshay Ghodake, Ayushi Tiwari · DevConf.CZ 2025
• 45mFast and Safe Production Monitoring of JVM Applications with BPF MagicSasha Goldshtein · LISA17
• 35mOpening the Box: Diagnosing Operating-System Task-Scheduler...Julia Lawall · SREcon24 Europe/Middle East/Africa
• 55mPerformance tuning of Red Hat Enterprise LinuxLarry Woodman · DevConf.CZ 2019
• 21mGraphing tools for scheduler tracingJulia Lawall · Linux Plumbers Conference 2023