Diagnostic Tooling - Be Ready when Things Go Wrong in Production
summary
Diagnostic working group catalogs gaps, pushes node-report toward core, defines support tiers and CI, and lands v8 fixes to restore flame graphs visibility. Demos show node-report capturing OOM with command line and heap details, perf + flame graphs finding a hot toString cost caused by unused event emission, and llnode plus core dumps enabling heap inspection. Work also advances async_hooks and integrates trace events for unified tracing.
AI-written summary · about
chapters (9)
- 0:00Start
- 2:13Fail Fast?
- 4:24Scenarios
- 5:34Toolkit
- 15:43Some Examples
- 16:54Node Report
- 21:46Scenario 2 - What is my app doing
- 25:32Scenario 3 - Flame Graphs
- 26:53Recap
related talks
25mNode js applications diagnostics under the hood
16mNode.js Diagnostic Best Practices
23mThrow Me a Lifebuoy: Debugging Node.js in Production with Diagnostic Reports
50mDebugging Node.js in Production
57mTroubleshooting Node.js Applications Under the Hood
44mNow You See Me Too: Visual Tooling for Advanced System Analysis
50mPost-Mortem Debugging in Node.js
52mApplication-Level Debugging and Profiling: Gaps in the Tool Ecosystem