Announcing MySQL Plugin for Oracle Enterprise Manager 12c Cloud Control

Posted in: Technical Track

MySQL management plugin for EM 12c has been long overdue. I’ve initially migrated the older plugin to EM 12c about 6 months ago, and few dozen people received this as the initial beta of the plugin. It worked OK but didn’t use any of the new 12c features, and its home page was a bit of a mess in the EM 12c Cloud Control web interface.

I’ve had lots of new features to add, but I didn’t really have much time to invest into completing them all. Finally, I decided to just finish the home page dashboard and clean it up from all unfinished new features. I did, however, finish MySQL Slave configuration and status monitoring, which was the largest gap in the the functionality of the previous plugin. There is no custom UI for MySQL slave management yet (that’s coming), but monitoring is available as standard metrics and configuration management features.

The home page looks like that in the released plugin version

There, you can see the summary of the MySQL instance configuration in the top left corner. Below that is the availability history, and to the right are several performance charts showing you the current snapshot of MySQL instance and how it’s doing. At the bottom of the page, there is a standard EM incident management panel.

Note the standard plugin menu in the top-left corner below the target name:

This is where you can access standard EM 12c user interface to metrics, incidents, configuration management, reports, and target setup.

For example, the Configuration section gives you access to the latest target configuration as well as the history and comparison with past or another MySQL target configuration:

Note that Slave Configuration is new compared to the previous plugin version for Grid Control. The configuration is obtained from show slave status. Show slave status also has some real-time setting, which I collect at Slave Status Realtime. Most of these settings are only important in the current snapshot, so by default I didn’t schedule the collection of those metrics. They include the position in bin logs and relay logs, for example, as well latest errors of SQL and IO threads. A few metrics are also collected in Slave Status collection group — Slave Lag in seconds (using Seconds_behind_master of show slave status) and the state IO and SQL slave threads, which might be important to look at, and set thresholds to raise incidents automatically.

Known issues:

  1. The plugin home page generates run-time errors if MySQL target is down. This is because the dashboard is using real-time data, and it can’t retrieve it. I’m looking into ways to solve this. It’s a bit annoying as there are multiple errors popping up, but you don’t need to close them and you can just select page from the menu that you want (like target setup or the latest configuration). You don’t really have anything useful on the home page when target is offline.
  2. If you’ve been using the previous private beta version, please undeploy it and deploy the new plugin from scratch. The new Plugin Id is different, but the target type is the same so you will encounter an error during deployment. Note that you might loose target instances when you undeploy the old beta release of the plugin along with its configuration and metrics history. When I tested it on my setup, I was able to undeploy and deploy a new one without removing existing targets. All their settings persisted, but to be on the safe side I would remove old targets and add them again. The future plugin upgrade should be straightforward.

Please report any other issues you encounter — I’m sure there will be plenty as I haven’t done extensive testing. Consider this more of the first release candidate quality. Please also comment if you get it to work along with you EM detailed version for OMS and Agents, Agent OS, MySQL OS, and MySQL version.

Useful links:

Happy monitoring!

Want to talk with an expert? Schedule a call with our team to get the conversation started.

About the Author

What does it take to be chief technology officer at a company of technology experts? Experience. Imagination. Passion. Alex Gorbachev has all three. He’s played a key role in taking the company global, having set up Pythian’s Asia Pacific operations. Today, the CTO office is an incubator of new services and technologies – a mini-startup inside Pythian. Most recently, Alex built a Big Data Engineering services team and established a Data Science practice. Highly sought after for his deep expertise and interest in emerging trends, Alex routinely speaks at industry events as a member of the OakTable.

8 Comments. Leave new

Quite Release MySQL Plugin — bug fixes | The Pythian Blog
December 11, 2012 1:30 pm

[…] known issues exist as for release. Useful […]


Hi Alex,

When is the next version of MySQL Plugin for OEM coming out? Badly need Galera cluster monitoring in OEM.


Hi Alex,
We have a Proof Of Concept the node multimaster cluster. One of the instances gives the message “The value of Binlog_disk_pct is 100”. I have checked the disksize. It’s OK now. I also don’t see that binary log is turned on other than the necessary binlog_format=ROW.

All of the 3 instances also report “The value of Slave_SQL_Running_flag is 0”. But I don’t have any slaves. What can I do?
Thanks in advance.

Alex Gorbachev
April 6, 2017 7:35 am

Hi Mariane.

The pluging doesn’t collect any disk data – it’s a job of your OEM host target to monitor this. This percentage indicator is about the percent of binlog writes going directly to disk rather than memory (binlog_disk vs binlog_memory). If binary logging not turned on that this won’t make sense and you can remove alerting thresholds on this metric from OEM. Default thresholds are always just samples.

If you don’t have any slaves then you don’t need to monitor that metric so you can remove the alerting thresholds. However, I’m not sure how you have multi-master replication cluster without any slaves (unless you are using some other type of replication).

In short, the plugin is a tools for an experienced MySQL DBA who is expected to fine tune the monitoring alerts based on specifics of the deployment.

Hope this helps.


Hi Alex,
Thanks for your quick response.
I am new to mysql. A collegue of mine set up the galera cluster and I installed the plugin. I will have to check te configuration to see how the data is replicated to the other nodes.


I checked the configuration file and the parameter wsrep_sst_method is set to rsync.
So should I still have binary logs turned on if the other nodes are kept in sync with rsync?
I thought that it wasn’t necessary. If so We will have to evaluate our current parameters in the config file.
Tkans in advance.


Hi Alex,
the MySQL Plugin cannot be rolled out with 13cR2 agents.
Do you have some plans when this will be supported ?


looking the same…


Leave a Reply

Your email address will not be published. Required fields are marked *