Slow Query Log Analyzer

The Slow Query Log Analyzer assistant helps database teams move from vague performance complaints to a precise, prioritized list of the queries most worth optimizing. Slow query logs and query statistics views are the starting point for almost every serious database performance investigation — but turning raw log data into an actionable optimization plan requires systematic analysis and the right interpretive framework.

This assistant guides you through the process of working with slow query log output from any major database. For MySQL, it helps you interpret the slow query log format and work with tools like pt-query-digest to aggregate and rank queries by total execution time, call frequency, and per-execution cost. For PostgreSQL, it helps you query pg_stat_statements to surface the queries with the highest total time, worst planning overhead, and most variable execution duration. For SQL Server, it covers the Query Store views that reveal plan regressions and top resource consumers.

From this log data, the assistant helps you build a prioritized optimization backlog. Not all slow queries deserve equal attention: a query that runs once per day for 10 seconds is less urgent than one that runs 50,000 times per day for 50 milliseconds each. The assistant helps you calculate total impact, distinguish between queries that are slow because of structural problems versus those that are slow because they process large data volumes inherently, and identify patterns — such as a family of queries that all share a missing index — that suggest a single fix with broad impact.

The assistant also helps you set up and configure slow query logging correctly: choosing appropriate thresholds, ensuring the log captures the information needed for analysis, and establishing a workflow for regular log review as a proactive performance practice rather than a reactive firefighting tool.

Ideal users include DBAs responsible for database health across a busy application, platform engineers building database observability pipelines, and development teams investigating latency issues reported by users or monitoring systems. This assistant turns slow query data from noise into a clear signal.