Build resilient systems at scale
May 27–29, 2015 • Santa Clara, CA

Box Weather Station: Open source tools for MySQL performance and monitoring

Gavin Towey (Box, Inc)
4:10pm–4:50pm Friday, 05/29/2015
Location: Mission City M1-2

Prerequisite Knowledge

Attendees should know the basics of installing and running MySQL databases in production. In addition knowledge of how to set up apache with php will be necessary to run these tools.


We present two open source tools developed at Box that are simple to use and very powerful:

Anemometer is a query performance tracking tool that leverages Percona’s pt-query-digest tool to help gather detailed query execution data, and then provides a drop-dead simple interface for analyzing and reporting on that data. It can show you aggregate views across all your database servers, which makes finding your applications hotspots easy, even when you have large distributed database environments. It also exposes a simple API, which makes powerful DevOps integrations possible such as creating automated monitoring reports, and automatic query checking in your integration pipeline.

Raingauge is a tool for capturing forensic data from your databases when the unexpected occurs. It’s built on top of Percona’s pt-stalk, which it uses to collect the data — but it takes it to the next level by aggregating your collections from all your servers in one place. You can see trends at a glance, and explore the data from its simple web interface. This tool is critical for minimizing outages and failures — you will never again miss important data you need to resolve issues. You can collect the data you need the first time, and move on to fixing problems instead of waiting to reproduce sporadic faults.

Photo of Gavin Towey

Gavin Towey

Box, Inc

Gavin Towey is a DBA at working on database operations, performance, and writing tools to make managing databases easier. He has worked with MySQL since 3.23 as a developer first, and has been a dedicated DBA for the past 5 years, and has managed mysql installations with hundreds of instances.