n-plus-one-detector

The N+1 query problem occurs when an ORM-backed endpoint fires one parent query followed by dozens of individual child lookups. Each query looks fast in isolation, but total request latency scales linearly with the result-set size, creating severe performance bottlenecks.

Fix N+1 queries by applying scoped eager loading, batched loaders, or grouped aggregates depending on the access shape. The Skill diagnoses the exact chatty pattern using query counts and request traces, then recommends the smallest targeted fix for your specific data-access loop.

Prevent N+1 regressions by designing query-count budget tests that assert the maximum allowed database requests per endpoint. These regression guards monitor request-scoped query evidence, ensuring the same chatty data-access defect cannot silently return during future development.

Yes, chatty query detection applies directly to GraphQL APIs where resolvers fire one database request per field across a list of items. The Skill pattern-matches fixes like batched loaders and scoped eager loading specifically for resolver loops and serialized list endpoints.

Diagnosing chatty query performance requires per-request query evidence such as query counts, ORM instrumentation logs, and request traces. This evidence-first approach confirms the exact data-access pattern before proposing scoped batching, eager loading, or denormalization changes.

Use request-scoped caching only when appropriate for repeated identical lookups within a single request lifecycle. For list endpoints where latency scales with result-set size, scoped eager loading or batched loaders are generally preferred over caching to eliminate serial awaited calls.