Cosmosdb Best Practices

If you're building on Azure Cosmos DB, this is a comprehensive rulebook covering data modeling, partition key design, query optimization, and SDK patterns across 100+ specific guidelines. It's built for code review and refactoring, with rules prioritized by performance impact and organized into 11 categories from critical (partition keys, data modeling) to medium priority (monitoring, tooling). The guidelines are concrete: use hierarchical partition keys ordered broad to narrow, avoid cross-partition queries, never reuse request options objects in the Java SDK, serialize enums as strings. It covers vector search and RAG patterns too. Think of it as the distilled wisdom from running Cosmos DB at scale, formatted for Claude to apply during code generation or review.