Performance is about more than time series charts
summary
Argues that fixing the slowest metrics wastes effort and recommends prioritizing pages by user impact: an x‑axis plot with green dots for response times and tall blue bars showing pages users care about reveals three high‑impact pages to fix first. Uses treemap to show both metric and underlying scale, histograms for third‑party aggregate effects, and a dimension co‑occurrence ring to find intersections (eg slow in California when served from a specific datacenter). Overlays forecasts (Holt‑Winters / stddev) to enable adaptive alerts and reduce alert fatigue.
AI-written summary · about
chapters (6)
- 0:00Intro
- 0:15Time series charts
- 0:58Waterfall charts
- 1:56Root cause analysis
- 3:23Thirdparty analysis
- 4:24Dimension cooccurrence
related talks
38mThe Structure and Interpretation of Graphs
26mImproving How We Observe Our Observability Data: Techniques for SREs
36mMoneyball for Performance Metrics
36mHunting the Unicorn
12mPerformance In Context
10mTurning data into leverage
48mHow I learned to stop worrying and love UX metrics
48mRevisiting Performance Budgets