I am a great fan of “MONyog”(agent-less monitoring tool) and am using it for a long time now.

I feel it’s a great tool and has saved many sleepless nights of mine.

Its “Error Log Monitoring” feature monitors the MySQL Error logs and sends alerts through mails or SNMP traps for the errors that are logged in.

Cheers for MONyog…

Other tests for individual things which have been seen as problems exist:

- Memory usage
- Long-running transactions (seen in SHOW ENGINE INNODB STATUS)
- Queries taking too long (although the slow log is generally enabled)
- MyISAM tables getting nearly full

We also monitor replication, to check that it’s up to date AND that the data are in sync (using mk-table-checksum). We check the schema are in sync too.

This all feeds back into our Hobbit monitoring system which sends alerts to an on-call engineer who is available 24/7

We run a lot of mysql servers (100s), in different roles in the application – not all of them require all the monitors above but they all get some monitoring.

Additionally the web application (which connects to some of the databases) is monitored both by our Hobbit and several third party testing companies – these also alert the on-call engineer.

I use monyog, a MySQL and Linux monitoring tool from Webyog (http://www.webyog.com/). Well, actually, I have been using monyog for a couple of years now, and I’m quite satisfied with the software.

The best part of monyog is that it is agent-less. Having to install nothing additional on the server machine saves a lot of hassle.

I’ve configured MONyog to check on my server every minute… and in case there’s any trouble, I get alerts in my mailbox. Monyog also monitors OS metrics, and the web-based interface provides information in a graphical format which is very easy to interpret. Monyog handles my SSH keyed authentication tests as well.