Query Performance Analyst

Slow queries are one of the most common and costly problems in database-driven applications. A query that takes seconds instead of milliseconds can cascade into application timeouts, degraded user experience, and overloaded database servers. This AI assistant is built for database administrators, backend developers, and data engineers who need to identify, understand, and fix slow query performance across relational database systems.

The assistant helps you analyze SQL queries from multiple angles. It reads and explains execution plans — the internal roadmap a database engine uses to process a query — translating cryptic operator names, cost estimates, and row count predictions into plain language that tells you exactly where time and resources are being spent. Whether you're working with PostgreSQL's EXPLAIN ANALYZE output, SQL Server's graphical execution plans, or MySQL's EXPLAIN FORMAT=JSON, the assistant walks you through the key bottlenecks.

Beyond reading plans, the assistant helps you rewrite queries for better performance — identifying inefficient patterns like correlated subqueries, excessive function calls in WHERE clauses, unnecessary DISTINCT operations, or poorly structured JOINs. It recommends index strategies based on query patterns: when a composite index would help, when a partial index is more appropriate, and when over-indexing is creating write overhead without read benefit.

The assistant also helps you understand the interaction between query structure and database statistics — why the optimizer sometimes chooses a bad plan, how stale statistics lead to poor cardinality estimates, and when manual hints or plan guides are warranted. It covers both OLTP query tuning (fast, frequent, small-result queries) and analytical query optimization (large scans, aggregations, window functions).

This assistant is ideal for backend developers debugging production slowdowns, DBAs handling query escalations from application teams, and data engineers optimizing ETL pipelines that run against relational databases.