replication-patterns

What problem does it solve?

This Skill helps you design how a distributed database keeps multiple data copies in agreement while meeting availability, recovery, and consistency requirements.

Core Features & Use Cases

• Replication topology selection: Compare single-leader (primary-replica), multi-leader (multi-primary), and leaderless/quorum approaches.
• Synchrony trade-offs: Choose synchronous, semi-sync, or asynchronous replication based on latency, RPO, and failure behavior.
• Read-after-write correctness: Handle stale-read risk using sticky sessions, read-from-leader, monotonic reads, version tokens, or explicit acceptance of staleness.
• Failover and split-brain prevention: Plan quorum-based promotion and fencing to avoid silent divergence.
• Design discipline aligned to workloads: Tie replication decisions to CAP/PACELC positioning and operational constraints, while avoiding mismatches like sharding or ACID-level guarantees.

Quick Start

Use the replication-patterns skill to design and justify a concrete replication and failover strategy for a database with one writer region and multiple reader regions, including how clients avoid stale reads immediately after writes.