This page is for viewing and controlling the MySQL instances that are being monitored by MySQL Enterprise Monitor, and configuring MySQL instances that will be monitored.
Keep in mind that the Status Summary at the top right of MySQL Enterprise Monitor User Interface is directly related with the information shown on the MySQL Instances page.
This section contains three buttons that let you:
Create Group: Creates a MySQL instance group that will immediately be displayed on the MySQL Instance Details listing.
Add MySQL Instance: Allows you to add a MySQL instance for MySQL Enterprise Monitor to monitor.
Add Bulk MySQL Instances: Allows you to add multiple MySQL instances for MySQL Enterprise Monitor to monitor, all of which use the same configuration details.
Connections for every Agent may be configured from this dashboard,
when the Agent is installed with the Installer, or with the
agent.sh script that is bundled with each
The unmonitored MySQL instances section is only shown when both the
MySQL Process Discovery advisor is active, and
there are unmonitored MySQL instances that where detected by this
This feature needs to know the port number, basedir, and datadir to
know that a connected MySQL server matches one that is unmonitored.
mysqld does not return enough information to
do this any other way, so essentially unmonitored MySQL instances
are found if MySQL is started by
A MySQL instance will be seen as "unmonitored" if a local Agent discovers a MySQL instance that it is not monitoring. This MySQL instance could be seen as both monitored and unmonitored if a different Agent is monitoring it, or if it is being monitored remotely.
Workarounds including using Ignore Instance on the "unmonitored" MySQL instance, or to override the Advisor to not to perform process discovery.
From the list of unmonitored connections, you can:
Monitor Instance: Only used on processes that are not currently attempting a connection. This opens the Add MySQL Instance menu, with pre-populated data filled in from the discovery process.
Ignore Instance: Removes the discovered instance from this list. Useful for hiding MySQL instances that you are not interested in monitoring.
Cancel Pending Connection: Can only be used for connections that are currently being attempted, but were not yet completed. This will effectively cancel the attempt to establish a connection to that instance.
After adding a connection to an unmonitored process, the MySQL instance may appear as both unmonitored and monitored for a brief moment of time.
You can perform Monitor Instance, Ignore Instance, and Cancel Pending Connection actions on a single discovered instance, or when configuring multiple (bulk) instances.
When you are monitoring multiple discovered processes in a bulk operation, the menu will not show an Agent selection, nor will it show information for the connection itself. In other words, it will only show the credentials as seen in the screenshot below. Both the Agent and the connection details will automatically be selected on a per-instance basis based on the information extracted from the discovered process. The defined credentials are used for every instance of the bulk operation:
The bad connections section is displayed if there are bad connections to the monitored MySQL instances. Possible scenarios that cause bad connections are:
A connection that was created with the wrong credentials.
A connection that was issued to an unknown host.
Any connection that could not be established.
You can get more detailed information on the error that caused the bad connection by hovering your mouse over it:
Usually a bad connection means that MySQL Enterprise Monitor cannot find any monitoring data from the MySQL instance, as MySQL Enterprise Monitor cannot access the instance, but this is not always true. For example, for a connection that is using less privileged users, the administrator credentials may be correct while the less privileged credentials may be wrong. In this scenario we could still monitor the server using the administrator connection while showing the bad connection for the less privileged users.
The screenshot below shows that a MySQL instance connection to 127.0.0.1:5168 is showing as a bad connection, but also notice that the MySQL instance is still being listed as correctly monitored.
The bad connections table also allows you to access the context menu for each of the bad connections, which allows the following:
2.x Agent connections cannot be edited from this interface.
Edit Connection: Generates a window to edit the connection details.
Delete Connection: Removes the connection. If you delete the connection, then it will no longer show as a bad connection, and it will be forgotten.
Keep in mind that if you remove a bad connection that is related with an instance (for example, a connection that is only partially bad or that was once good but it now shows as a bad connection) then you will leave that monitored instance connectionless which means that it will show highlighted in red on the monitored instances list below:
For each bad connection, MySQL Enterprise Monitor will actively try to establish that connection for a fixed period of time (this connection timeout defaults to 5 minutes). For example, let us say a bad connection was created due to bad credentials, and you changed the credentials using a standard MySQL client (outside of MEM) to match the credentials you set in MEM. In this scenario, the bad connection would disappear because it would become a good connection. After the timeout, MEM will stop making attempts to establish the connection, and information would be shown as:
The Unreachable Agents section displays information about the unreachable Agents, and provides a Delete Agent action.
This section is only visible if there are unreachable Agents
The MySQL Instance Details section shows the list of monitored instances. From this section you can:
Remotely monitored hosts and 2.3 Agents are highlighted with a caution icon ( ). Hover over this icon and review the suggested change.
Filter: This searches through the specific instances using keywords.
Delete Instance: This will remove this instance from the list of MySQL Enterprise Monitor's monitored instances.
Ignore Instance: This is similar to the Delete Instance operation, (since the instance will no longer be monitored), but ignored instance information is saved, and can be viewed and added later on. The connection properties are preserved.
Show Instance: Allows you to recover ignored instances.
To recover (show) an ignored MySQL instance, first find the
ignored instances by executing a search with the
Ignored Instance filter paramater set to
enabled. Select the ignored instances you
would like to show again, and then click the Show
Edit Instance: Allows you to edit any information concerning that instance.
Export Overview as CSV: Exports the list of monitored instances as a comma-separated values (CSV) file.
Control Groups: Accessed from the context menu on the group.
Add to Group: Adds a MySQL instance to the group.
Delete Group: Removes the group (the instances within the group will not be removed, but will be dissociated from the removed group).
Rename Group: Renames the group.
The text is captured as is in the text field. For example, HTML entities are not converted.
You can execute the Delete Instance, Ignore Instance, Show Instance, and Edit Instance actions on a single instance, or on multiple instances using the bulk operation. When editing multiple instances, the same logic used for monitoring instances applies.
When MySQL instances are ignored or filtered, the number of ignored instances is displayed:
And ignored MySQL instances are grayed out and italicized when listing them using the search filter: