performance-reviewer

You can audit code diffs in pull requests to detect N+1 queries and unbounded memory growth by automatically triggering reviews when changes touch database query logic or loop-heavy data transformations.

A performance review for production-scale code changes identifies runtime scalability defects like missing pagination, hot-path allocations, and blocking I/O in async contexts that only manifest under real production load.

This performance review tool does not trigger for all pull requests, but applies conditional automatic activation only when code diffs modify database query logic, caching implementations, loop-based data transforms, or I/O handling paths.

To prevent out-of-memory errors when adding a data export endpoint, audit the code changes to flag if the endpoint loads all records without pagination, catching unbounded memory growth before deployment.

To filter out premature optimization suggestions during a scalability audit, use confidence-calibrated findings that only report high-impact runtime issues with a confidence score of 0.60 or higher to reduce noise.

The limitations of automated performance reviews for async I/O handling include focusing strictly on high-confidence, production-observable issues like blocking I/O, and filtering out speculative runtime optimization suggestions.