Introducing a New Performance Monitoring Daemon
summary
A lightweight Python daemon called guider monitors CPU, memory, IO, sockets and function calls and automatically generates a text/SVG analysis report and runs pre-registered commands when thresholds are hit. It ships as a single script (~1 MB), implements ptrace/backtrace tracing for C/C++/Rust and Python stack sampling, exposes 135 commands and options, and keeps storage by pruning old reports. Runtime cost is small (about 1–5% CPU on one core), works on Linux/Android (limited Mac/Windows), and requires no extra kernel patches or external binaries except optional visualization libs.
AI-written summary · about
chapters (25)
- 0:00Introduction
- 0:13Summary
- 0:34Performance Factors
- 6:54System Resources
- 8:42Test Resources
- 9:34Function Calls
- 10:35Profiler
- 10:51Files
- 11:27Native Functions
- 12:51Python Function
- 13:30Cisco Function
- 13:54Signal Tracing
- 14:22Resource Graph
- 15:13Scheduling
- 17:45Problem Analysis
- 20:35Occurrence Condition
- 21:42Guider
- 22:12Commands
- 25:17Reports
- 26:52Report File
- 28:02System Resource Table
- 29:54Most Specific Information
- 31:00Event Command
- 32:07Conclusion
- 32:55Features
related talks
39mGuider: A New Performance Monitoring Daemon
53mGuider: Performance Tracing Features for Linux
37mFrom Continuous Tracing To Automated Insights: AI-Driven Performance Analysis With Guider
20mGuider: Lightweight Real-Time Performance & Fault Monitoring Framework for Embedded...
23mThe Hitchhiker's Guide to Advanced Python Monitoring
28mKnocking On A Process’s Door: Introspecting Live Python Processes
01h 00mInvestigating C++ Applications in Production on Linux and Windows
21mFrom PyPerf to py-spy - Everything You Need to Know About Python Profilers