Get to grips with the Basics
Fundamental Server Resources
Monitoring of the underlying server’s resources is the basis of MySQL database monitoring. Server CPU and memory usage, free disk space, free swap space, and network usage are all important metrics and the perfect starting point for troubleshooting in case of problems. Anturis will let you know in advance about any overloads of these resources, giving you the chance to prevent negative impact on database operations. You will also be able to spot usage trends over time so as to get insights on when you need to replace/upgrade the hardware serving the MySQL database.
The next obvious thing to know is MySQL availability. Failure of MySQL usually makes reliant applications completely inoperable. Anturis alerts you immediately in case MySQL itself or the underlying server goes down.
MySQL Error Log
MySQL error log contains all the errors occurring while MySQL is running as well as some context information, like stack traces. Anturis will watch the log file for error entries and keep you informed when it finds them.
Key MySQL Server Metrics
There are two ways in which MySQL metrics are useful to you: error detection and trending. In the first case you are notified of hazardous conditions that require your attention. Such conditions are defined by the thresholds on the selected metrics. In the second case, you have access to the metric graphs and you are able to troubleshoot by looking for unexpected changes and unusual patterns, and analyzing their reasons.
Here are a few examples of MySQL key performance metrics available with Anturis MySQL monitor.
- Slow query rate. Slow queries are eating the precious resources of the server, leading to degraded performance and application failures. Having a slow query rate that is too high indicates that it is a good time to look into the slow query log and check what’s going on.
- Connections usage. If your MySQL server reaches the maximum number of allowed connections the new client connections will simply be refused. Running out of connections signals an abnormal situation, such as when configuration changes may be needed.
- InnoDB statistics. This includes InnoDB buffer pool usage, its’ miss and wait rates, and the InnoDB log cache wait rate. Using this data you are able to understand whether or not InnoDB performance is optimal.
- MySQL query and key cache statistics. A low cache hit rate signifies that cache configuration is misaligned with application’s needs.