Zabbix Templates for MySQL
We recently had a client come to us, and ask us to improve their MySQL monitoring in Zabbix. So, we did. The approach we took was to port the MySQL script from the superb mysql-cacti-templates project to work with Zabbix. This works out well, because Zabbix is like cacti and nagios combined, and, what we wound up getting, are some templates that can alert us when InnoDB uncheckpointed bytes starts climbing rapidly.
In addition to the above benefit, you also get every MySQL graph that comes from the mysql-cacti-templates project making Zabbix with appaloosa-zabbix-templates* a first-class replacement for cacti when it comes to MySQL trending.
Finally, the same client is also generously donating time to create templates for Memcached/Membase and Bind9. So, look forward to more out of this project as we go forward into next year.
* Did you notice we really like our horses, here at PalominoDB?
Archives
- May 2013
- March 2013
- February 2013
- January 2013
- December 2012
- November 2012
- September 2012
- August 2012
- July 2012
- June 2012
- May 2012
- April 2012
- March 2012
- February 2012
- January 2012
- December 2011
- November 2011
- October 2011
- September 2011
- August 2011
- July 2011
- June 2011
- May 2011
- April 2011
- March 2011
- February 2011
- January 2011
- December 2010
- November 2010
- October 2010
- September 2010
- November 2009
- March 2008
- November 2007
- October 2007
Comments
Hey LinuxFood,Thanks a lot for your contribution. I am going to use you Zabbix templates.Cheers.
The answer is of course, "it depends", and a whitepaper would be required to really answer it well. When you say monitoring, do you mean monitoring for trend and workload analysis, or for fault detection and alerting? \
Generally for either, one will want to track basic OS stats - CPU consumption, load average, memory utilization, swapping usage and swapping activity, IO activity for disk (response time, queues, throughput, distribution of reads to writes etc...) as well as network statistics.
Next step is MySQL specific - and this can vary but basics include replication lag, counter statistics (how many selects, inserts, updates, deletes), innodb specific counters, sorts, select handlers (range scans vs. key lookups vs. full table scans), usage of your query cache, myisam key buffer and innodb buffer pool...
There's a lot out there. For trending and workload analysis I prefer to cast a wide net, as you never know where a bottleneck or the symptoms of a problem might manifest. For alerting, you need to decide what is going to impact your service and alert accordingly - http://dev.palominodb.com/docs/nagios_plugins_2011_04.pdf
Hi
Thanks a lot for this topic.
May anyone give me some guidance about important parameters in mysql monitoring?in fact,which parameters is important to focus ?
Thanks in advance
Yes, there are screenshots at the Google Code page that was linked in the blog post. http://code.google.com/p/appaloosa-zabbix-templates/
any shiny screenshots to stare at ? :)
That's great. I never had a customer willing to sponsor this port but always hoped it would happen.
In case it's not clear, this project is 100% free, open source, and accepts community contribution.
Reply