Statistical Methods and Tools for System Performance Analysis
summary
Describes a method for performance testing when only real, highly seasonal traffic can be applied: identify input and output metrics, track them across periods, and compare metric-to-metric relationships to assess headroom, overload behavior, and to compare two releases on the same bench. Mentions using open-source tools ipython notebook, Graphite, Diamond, pandas and scikit-learn for analysis and anomaly detection.
AI-written summary · about
chapters (11)
- 0:00Introduction
- 1:32What We Want to Know
- 7:10Business Metrics
- 7:53System Metrics
- 8:14What Can Be Seen
- 9:34So Many Graphs
- 11:25Looking for Correlations in Metrics
- 17:26Dependencies of Filters
- 20:19Linear Regression
- 23:52Metrics and Graphs
- 26:09Visualization
related talks
47mPractical Performance Analysis
38mPerformance 101
35mHonest Performance Testing
40mPerformance Testing Done Right
49mAll Levels of Performance Testing and Monitoring in Web Apps
50mFrom Load Testing to Reliability Engineering: Making Performance Testing Predict Production Behavior
01h 00mPerformance In Modern Times: An Evolutionary Journey
53mEarly Performance Testing: Finding News You Can Use