Innodb Pending Reads and Writes

This alert is triggered when the aggregate number of pending reads and writes in the MySQL buffer pool is greater than 2 for 5 minutes, which indicates the database engine is backlogged and waiting on resources.

Max Connection Errors per Second

This alert is triggered when there are greater than 1 errors against the max_connections limit in a 5 minute window, which indicates you have requests to your MySQL instance that are failing to connect. This setting's default is 501, but can vary based on the underlying resources available to your instance. You can review your current max_connections limit with this query: SHOW VARIABLES LIKE 'max_connections';

Questions per Second

This alert is triggered when the current rate of Questions is greater than 2 standard deviations above the baseline for 60s, which could be an early indicator of a saturation problem for your instance. It is important to note that this alert is disabled by default and you need to edit the configuration in New Relic One to add a targeted MySQL instance: "WHERE displayName = 'MySql Instance Name'" This allows the baseline to be calculated against a single instance instead of all running MySQL instances being monitored.

Slow Queries per Second

This alert is triggered when the number of slow queries per second is greater than 5 for 5 minutes, which could indicate capacity issues or a query that has been changed and is experiencing performance issues. The Slow_queries counter increments based on your settings applied to MySQL's long_query_time parameter (default 10s), which you can review with this query: SHOW VARIABLES LIKE 'long_query_time';