Design HA Database Architecture

Analyze and document the architectural tradeoffs between using a replicated database shared by multiple Gateway replicas versus having each replica maintain its own local database with a single writer feeding data. A shared PostgreSQL database enables multiple Gateway replicas to serve queries from the same data source, providing high availability and load balancing for read queries. However, it introduces complexity around concurrent writes, requires database replication and failover mechanisms, and creates a potential single point of failure. Alternatively, each Gateway replica could maintain its own local SQLite database with only one replica actively synchronizing blocks, providing simpler architecture but requiring client-side load balancing and handling of stale reads. Document the pros and cons of each approach, recommend a solution based on production requirements, and design the implementation accordingly.

---

**Quarter:** Q3

The issue names no files, tests, or entry points. Start by mapping the Gateway replicas, block synchronization, and database interactions, then compare shared PostgreSQL with per-replica SQLite against production requirements; done means a documented tradeoff analysis, recommendation, and implementation design.

Written by the indexing model from the issue text.