How to monitor your database

Baron Schwartz demonstrates how to monitor a database by understanding the difference between workload and resource monitoringand the golden signals for each.

The first time Baron Schwartz tried to monitor a database, he was overwhelmed. There were hundreds of variables to monitor as well as lots of Nagios check scripts that each had dozens of checks. He wasn’t sure what alerts to set up, so he set up too many and got a bunch of noise as a result. A couple of years later, Baron returned to that company and found all those alerts still in place, still spamming everyone, but the company had just filtered every alert to the trash. Join Baron to discover how he learned to do monitoring better, so you won’t make the same mistakes he did. Any sophisticated system like a database has many more instrumentation points than you should actively monitor. The trick is approaching it with a sound monitoring framework in mind. Baron shares the framework he’s developed over many years, which breaks monitoring into a holistic approach that’s easy to understand and makes it obvious what kinds of data are useful for what purposes. You’ll learn the seven golden signals (yes, seven and not four), how workload and resource performance are complementary and necessary for a complete understanding of database health and performance, and how to monitor technology-specific “sharp edges.” You’ll also see how these concepts work with a few databases, including MySQL, PostgreSQL, and MongoDB.