Continuous Profiling - the 4th Pillar of Observability
summary
Continuous profiling is presented as a fourth observability signal that runs always on with subpercent overhead, uses Java Flight Recorder and sampling plus true CPU profiling contributions to JFR, and lets teams slice profiles by span ID tagging or custom labels to correlate traces and stacks. Real deployments at Datadog report velocity and cost wins — one campaign saved $17M in two months and another freed ~19,900 cores saving about $3M — and OpenTelemetry is standardizing the profiling signal.
AI-written summary · about
chapters (19)
- 0:00Introduction
- 2:07Why profiling matters
- 6:21Traditional vs modern needs
- 8:28JVM history to profiling
- 11:18Flight Recorder and sampling
- 15:52Context in distributed systems
- 17:36Open sourcing JFR
- 19:22Solving unknown problems
- 20:04Profile production systems
- 21:30Profiling boosts velocity
- 23:57Cost savings with profiling
- 26:00Profiling and engineering culture
- 28:09Customer success stories
- 30:16Improving profiling tools
- 32:05Importance of profiling context
- 36:04Profiling as 4th pillar
- 37:27Java’s profiling tech
- 38:55Final thoughts
- 39:36Q&A
related talks
40mContinuous Profiling, the missing piece in your observability puzzle!
56mContinuous Profiling the missing piece in your observability puzzle
52mThe missing piece in your observability puzzle!
01h 13mContinuous Profiling: The Missing Piece in Your Observability Puzzle
24mCommunity Driven Breakout Session: Continuous Profiling
50mObservability Beyond the Three Pillars — Continuous Profiling With Alibaba Dragonwell
24mWhy Does Continuous Profiling Matter to Developers?
06mContinuous Profiling