How doe MySQL Monk works

Uppon first run, MySQL Monk creates a simple table called mysql_monk in the replicated database.

mysql> describe mysql_monk;
+-------------+----------+------+-----+---------+-------+
| Field       | Type     | Null | Key | Default | Extra |
+-------------+----------+------+-----+---------+-------+
| master_id   | int(11)  | NO   | PRI | NULL    |       |
| last_update | datetime | NO   |     | NULL    |       |
+-------------+----------+------+-----+---------+-------+

• For each master, MySQL Monk updates the timestamp in that table periodically (as defined in mysql_monk.monitor.update_interval).
• For each slave, it checks if if the timestamp on the slave is close enough to the timestamp on the master periodically (as defined in mysql_monk.monitor.check_interval)
• if the timestamp difference is greated than in mysql_monk.monitor.max_allowed_lag, and event is raised (which usually results in an email being sent).
• if the timestamp difference was greater than the max allowed lag, and the situation is resolved, another email is sent - indicating that all is well again.
• if a server cannot be contacted (crashed? just turned off?), and email is sent. if it comes back online, an email is sent.