TechPrufer
Get started
PlanetScale
PlanetScale logoPlanetScale · Feature trail

MySQL Memory Profiling

1 post · Apr 2024

What this trail covers

PlanetScale provides tools and techniques for profiling memory usage in MySQL queries. This includes leveraging the `performance_schema` to instrument various memory components and sampling memory usage at different granularities (global, by thread, by user, etc.). The post details how to identify memory-intensive operations like sorting and how to collect and visualize memory usage over time using custom scripts and libraries like Matplotlib. This capability helps users understand and optimize the memory footprint of their MySQL workloads.

All PlanetScale trails
  1. Profiling memory usage in MySQLThis post introduces and details methods for profiling memory usage in MySQL queries. It explains how to use `performance_schema.setup_instruments` and `performance_schema.memory_summary_by_thread_by_event_name` to monitor memory consumption at a thread level. The post provides SQL queries for sampling memory usage and presents Python scripts for collecting and visualizing this data over time, including a version with Matplotlib for live plotting. It highlights specific memory categories like `memory/sql/Filesort_buffer::sort_keys` and discusses the trade-offs and limitations of per-query memory tracking.