High Availability — Zero Downtime in Production

A production system that goes fully offline for a single server failure is not production-ready. This category documents how to design and operate systems that survive hardware failures, maintenance windows and traffic spikes.

---

What this category covers

Topic	Status
SPOF, RTO, RPO — concepts	✅
Nginx load balancer — upstream, algorithms	✅
Passive health checks	✅
Sticky sessions vs centralized sessions	✅
MariaDB master/slave replication	✅
Keepalived — VIP failover	⏸
HAProxy — frontend/backend, ACL	⬜
Galera Cluster — multi-master	⬜

Posts in this category cover Module 07 — High Availability and Module 21 — HAProxy of the roadmap.