Challenges in Measuring Erlang Application Performance
summary
Examines practical pitfalls in measuring performance of Erlang applications, focusing on latency measurement and high-speed counters. Covers problems in existing tools and pipelines—sampling, random-number generation, and distributed aggregation—illustrated with examples and numbers from an open-source project.
AI-written summary · about
chapters (13)
- 0:00Introduction
- 1:46Measurement challenges
- 2:52Measuring time intervals
- 5:18Data sampling
- 10:17Metrics libraries
- 15:15HDR histograms
- 17:20Threading issues
- 21:18Counting measurements
- 24:49Randomness in metrics
- 26:59Latency differences
- 28:44Cluster data aggregation
- 30:31Summary and recommendations
- 31:57Q&A
related talks
44mThink Outside the VM: Unobtrusive Performance Measurement
35mDiving into Erlang is a one-way ticket
25mGreen Benchmarking of BEAM Languages
01h 01mThe Performance Investigator's Field Guide
43mHow NOT to Measure Latency
42mMeasuring your Elixir Application
44mWhy Do Clients Nitpick?
38mTechniques and Tools for a Coherent Discussion about Performance in Complex...