MySQL Health Checks

Collection of Nagios compliant health checks for Automatron

Includes:

• available.py - Checks if MySQL is up and available by establishing connection
• status_metrics.py - Validates and compares values of defined metrics from "show status" SQL query

Installation

Simply copy the mysql/ directory into plugins/checks/.

$ cp -r mysql plugins/checks/

available.py

The available plugin is used to query MySQL's internal status system and determine whether the MySQL service is available or not.

Runbook Example

The below is an example of using the mysql/available health check in a runbook.

checks:
  mysql_up:
    execute_from: target
    type: plugin
    plugin: mysql/available.py
    args: --host=localhost --user=USERNAME --password=YOURPASSWORD

This plugin can be executed from either target or remote depending on the MySQL service's configuration.

Required arguments

The mysql/available plugin requires 3 arguments.

args: -s <mysql host> -u <mysql user> -p <mysql password>

If the show status query is unsuccessful or does not find the key it is looking for the check will return a CRITICAL status. There is no WARNING or UNKNOWN status for this check.

status_metrics.py

The status_metrics plugin is used to query MySQL's internal status system and alert when the defined metric exceeds the warning and critical thresholds.

Runbook Example

The below is an example of using the status_metrics health check in a runbook.

checks:
  status_metrics:
    execute_from: target
    type: plugin
    plugin: mysql/status_metrics.py
    args: --warn=20 --critical=10 --metric=slow_queries --host=localhost --user=USERNAME --password=YOURPASSWORD --type=greater

This plugin can be executed from either target or remote depending on the MySQL service's configuration.

Required arguments

The mysql/status_metrics plugin requires 7 arguments.

args: -w <warning value> -c <critical value> -t {greater, lesser} -m <metric> -s <mysql host> -u <mysql user> -p <mysql password>

The type flag is used to define whether or not the alert is triggered when the metric value is "greater" or "lesser" than the values defined as warn and critical.