Configuring
Measures for a Jmx Test
Java-based web servers and application servers (eg., Tomcat, JBoss, WebLogic, WebSphere, etc.) are common place in many IT infrastructures. Though eG Enterprise provides comprehensive monitoring models for most of these servers, some administrators might want to monitor certain additional server attributes that are not supported out-of-the-box by the eG agent. Likewise, in some other environments, custom Java applications may operate for which eG Enterprise may not provide out-of-the-box monitoring support.
To enable administrators to extend the capabilities of the eG agent to report on additional attributes related to Java-based web and application servers, and to easily develop custom monitoring models for unsupported Java applications, eG Enterprise offers the Jmx test type. Jmx (Java Management Extensions) is a Java technology that enables external programs like the eG agent to pull out performance metrics from a target Java application/server. By adding a test of type Jmx, you can easily equip the eG agent with the ability to leverage the JMX technology for connecting to any Java application/server, and reading the current values of those attributes of the target that interest you.
Note:
Prior to adding a new Jmx test, you need to enable the JMX support for the JRE of the target application. To know how, please refer to the eG Customization Manual.
This page enables you to configure the measures for a test of type Jmx. To access this page, click on the Configure button in the Measure Selection Option page in the eG administrative interface.
The procedure for configuring measures for a Jmx test may slightly vary depending upon the option chosen from the MEASURE SELECTION OPTIONS page. Discussed hereunder are the individual options offered by the Measure Selection Options page, and how your choice will impact measure configurations for a Jmx test.
MBean Auto Discovery or Upload MBean File
If you had chosen the MBean Auto Discovery option or the Upload MBean File option from the Measure Selection Options page, then follow the steps discussed below to configure a measure for the Jmx test:
-
The Measure index determines the order in which the measures are to be displayed in the monitor console. For example, if you select 1 as the Measure index then the corresponding measure will appear first in the list of measures displayed in the monitor console. Note that the same Measure index should not be assigned to two different measures of a particular test.
The Measure name field indicates the name of the measurement (this will be displayed in the eG monitor interface).
The Database column size field indicates the size of each database record corresponding to a measurement value. For example, Number(7,4) indicates that the output of the measurement will be a number in the range 0 to 100, with the fractional value being limited to four decimal places.
The Unit (whether %, seconds, requests/sec, etc.) specifies the unit in which the measurement's value is reported. The required unit can be selected from the options available in the list box.
In addition to the above, this test type requests the specification of a Process method. By selecting an option from the Process method list box, you can indicate the processing that must be performed on the script / batch file's output before passing the results to the eG agent. The options offered by this list box are:
| Option |
Information |
| UNALTERED
| Selecting this option will ensure that no additional information is displayed along with the specified measure in the monitor console. In other words, the Current Value of the measure is displayed.
|
| PERCENT_INCREASE
|
Selecting this option will display the percentage by which the current value of the measure exceeds its previous value (i.e. the value of the measure during the previous test execution). While an increase displays a positive value, a decrease will display a negative value. The formula used is:
[(Current Value − Previous Value) / Previous Value] * 100
|
| PERCENT_DECREASE
| This option will display the percentage by which the current value of the measure falls below its previous value. While a decrease displays a
positive value, an increase will display a negative value.
[(Previous Value − Current Value) / Previous Value] * 100
|
| PERCENT_CHANGE
| This displays the percentage change between the current value and the previous value.
ABS[(Current
Value − Previous Value) / Previous Value] * 100
|
| RATIO
| This displays the ratio of the current value of the measure over its previous value.
(Current Value / Previous Value)
|
| RATE
|
This
displays the result of the following formula:
(Current Value − Previous Value) / Time since the last measurement
|
| DIFFERENCE
| This
displays the absolute value of the difference between the current value
and the previous value.
ABS(Current
Value – Previous Value)
|
By default, Unaltered is selected as the Process Method.
-
Sometimes, administrators may want to convert the unit of measurement of a performance metric before displaying the same in the eG monitoring console. For instance, a test may extract an ‘MB’ value from the target server as a measure, but while reporting it in the eG monitoring console the administrators may want the unit of the measure to be changed to ‘KB’ instead. Similarly, a duration value originally available in ‘minutes’, may have to be changed to ‘Secs’ before it is displayed in the console. If you want the unit of the measure being added to be so converted at test run time, then select a Conversion Factor from the list. For example, if a value in ‘Bytes’ needs to be converted to ‘MB’, then select /1048576 (Bytes to MB) as the conversion factor. By default, 1 is chosen as the Conversion Factor; this implies that, by default, unit conversion does not take place for a new measure at test run time.
-
The alarm window in the eG monitor interface displays all the currently outstanding alarms. Along with each alarm, the list of tests corresponding to the alarms and a brief description of the problem are displayed. For example, a disk space alarm indicates that the test reporting the problem is DiskTest and the problem description is “High disk utilization {C:}”. The problem description corresponds to the measure - Pct_disk_util- that has violated its thresholds. In this example, the Alarm display string corresponding to the Pct_disk_util measurement is “High disk utilization”. The alarm display string rather than the measurement name is displayed in the alarm window so as to make it easier for IT infrastructure operators to understand and use the eG Enterprise suite. When configuring a measurement for a custom test, specify the text string that must be displayed in the eG alarm window when the corresponding measurement violates its threshold.
If the MBean Auto Discovery option or the Upload MBean File option is chosen from the Measure Selection Option page, then eG Enterprise will automatically discover the MBeans contained within the Domain chosen from that page, and will make them available for selection in the Mbean name list. From this list, choose the MBean that reports the values for the measure that is being added.
Upon selection of an Mbean name, all the attributes of the chosen Mbean will be automatically discovered and displayed as the options of the Attribute name list. Now, from the Attribute name list, select the Mbean attribute that reports values for the measure being added.
-
By
clicking on the Add button, the measurement can be added. Similarly, you can add multiple measurements for a Jmx test.
Note:
Once a Domain is chosen for a test, all the measures of that test should be based on the MBeans and attributes within that domain only.
Manual Entry
If you had chosen the Manual Entry option from the Measure Selection Options page, then follow the steps discussed below to configure a measure for the Jmx test:
-
The Measure index determines the order in which the measures are to be displayed in the monitor console. For example, if you select 1 as the Measure index then the corresponding measure will appear first in the list of measures displayed in the monitor console. Note that the same Measure index should not be assigned to two different measures of a particular test.
The Measure name field indicates the name of the measurement (this will be displayed in the eG monitor interface).
-
The Database column size field indicates the size of each database record corresponding to a measurement value. For example, Number(7,4) indicates that the output of the measurement will be a number in the range 0 to 100, with the fractional value being limited to four decimal places.
-
The
Unit (whether %, seconds, requests/sec, etc.) specifies the unit in which the measurement's value is reported. The required unit
can be selected from the options available in the list have already been discussed in the previous section.
-
Sometimes, administrators may want to convert the unit of measurement of a performance metric before displaying the same in the eG monitoring console. For instance, a test may extract an ‘MB’ value from the target server as a measure, but while reporting it in the eG monitoring console the administrators may want the unit of the measure to be changed to ‘KB’ instead. Similarly, a duration value originally available in ‘minutes’, may have to be changed to ‘Secs’ before it is displayed in the console. If you want the unit of the measure being added to be so converted at run time, then select a Conversion Factor from the list. For example, if a value in ‘Bytes’ needs to be converted to ‘MB’, then select /1048576 (Bytes to MB) as the conversion factor. By default, 1 is chosen as the Conversion Factor; this implies that, by default, unit conversion does not take place for a new measure at run time.
-
The alarm window in the eG monitor interface displays all the currently outstanding alarms. Along with each alarm, the list of tests corresponding to the alarms and a brief description of the problem are displayed. For example, a disk space alarm indicates that the test reporting the problem is DiskTest and the problem description is “High disk utilization {C:}”. The problem description corresponds to the measure - Pct_disk_util- that has violated its thresholds. In this example, the Alarm display string corresponding to the Pct_disk_util measurement is “High disk utilization”. The alarm display string rather than the measurement name is displayed in the alarm window so as to make it easier for IT infrastructure operators to understand and use the eG Enterprise suite. When configuring a measurement for a custom test, specify the text string that must be displayed in the eG alarm window when the corresponding measurement violates its threshold.
If the Manual entry option option is chosen from the Measure Selection Option page, then, you will have to manually specify the exact attribute that reports values for the measure being added, in the Attribute name text box.
-
By clicking on the Add button, the measurement can be added.
When a measurement is successfully added, a response page indicating the measurement and its details will appear. Previously added measurements can be deleted using the Delete button that appears beside every measurement added. Such measurements can also be modified by clicking the Modify button beside the measure name.
The Back button enables the administrator to go back to the previous screen.
|