Understanding minimal viable quality promotes craft
summary
Argues that setting component promises to the true minimum required improves product craft by avoiding wasted optimization; uses SLI/SLO math, five-minute quantiles and workload classification to decide targets, and cites a time‑series DB that aimed for 2 milliseconds but settles on 2.5 seconds for user-facing queries, saving effort and unlocking product work. Recommends 99 of five-minute periods and limits like three monthly violations, warns about bad math (miscomputed 99 led to unnecessary hires) and advocates cascading transparency.
AI-written summary · about
chapters (9)
- 0:00Start
- 1:51What Is Minimal Viable Product
- 3:01Minimum Viable Quality
- 5:58How Do You Define Objectives
- 12:39Burn Down Charts
- 19:44Premature Optimization
- 20:14Pre-Premature Optimization
- 32:59Abstraction Boundaries
- 37:43Is It Possible To Define Smart Slos for a Complex Product with a Lot of Components
