Module 10: Production Hardening

Anchor outcome Zero unplanned downtime in 14 months. 3 planned maintenance windows totaling 47 minutes. 0 post-deploy rollbacks after adopting the preflight checklist. Monitoring catches 100% of P1 incidents before customer contact — MTTD dropped from 3 hours to 4 minutes.

Production hardening is the discipline between “it works in staging” and “it won’t break at 2am when the owner is on the water.” Every item in the Anchor preflight checklist exists because of a real incident that happened before the checklist did. The checklist is the scar tissue made systematic.

The 4 lessons

Lesson 10.1

Pre-Cutover Preflight Discipline

The 12-point checklist that runs before every deploy, post-deploy verification sequence (5 checks in 5 minutes), and the pre-committed rollback threshold that removes sunk-cost pressure from the decision.

Lesson 10.2

Zero-Downtime Patterns

The expand-then-contract migration pattern for schema changes, batch operations with sleep intervals to prevent table locks, and feature flags for changes that can’t be cleanly rolled back.

Lesson 10.3

Cache Flush Layers

The 4 independent caches (PHP OPCache, WP object cache, full-page cache, CDN edge), flush order, the OPCache one-shot script pattern, version strings for browser cache, and the complete post-deploy flush sequence.

Lesson 10.4

Monitoring and Alerting

External uptime monitoring targets, synthetic form submission check via WP cron, PHP error log scanning with incremental file position tracking, and alert routing (phone for P1, email for P2).