Part 3 of a series reflecting on 35 years of coding experience, focusing on strategies to recover from production defects. Key approaches covered include: vertical slicing (fractal slices) to limit blast radius of fixes, Event Sourcing to protect data integrity and enable state replay after corruption, and telemetry to quickly diagnose what went wrong. The author notes that thin slices often need only a single test, and introduces 'lifetime-tests' that cover multiple slices to reduce test duplication.
Sort: