Management Interface of Process Platform

This topic gives a brief description of the Management Interface of Process Platform.

The management interface of Process Platform is implemented as an abstraction layer for JMX, which supports concepts such as Components, Settings, Info properties, Counters, Operations, Alerts, and Problems.

Components

The Components in Process Platform are the application connectors that are exposed as JMX Managed Beans.

Settings

Settings are the knobs used by the administrator to tune the performance and scalability of the system. Conceptually, settings are the properties of the application connectors that are exposed as read/write or write-only attributes. Some of the settings that the administrator can configure are connection pool size, LDAP Cache, thread pool size, and so on.

Settings are displayed under the Attributes node of MBean in JMX Management tool. See Managed Components of Process Platform for detailed description of settings of each component.

Settings are classified into following types:

Hot: Hot settings are implemented immediately after the change and do not require a restart of the service container.
Cold: Cold settings are implemented only after the restart of the service container.

Info Properties

Info properties are properties of managed objects that are only informational to the administrator. The info properties are exposed as read-only attributes and are available under the Attributes node of JMX Management tool. The info properties are exposed as simple attributes, that are displayed as info_<propertyName>. For example the status of a service container is displayed as info_status.

Counters

In the context of JMX, counters are exposed as read-only attributes and are available under the Attributes node of JMX Management tool. For a detailed decription of the counters, see JMX Counters.

The counters are exposed as complex type attributes, that are displayed as ctr_<counterName>. For example cacheSize is displayed as ctr_cacheSize. The statistics are exposed as ctrv_<counterName>_<statistic>.

The counter types and the statistics for each counter type (xxxx is the placeholder for the counter name) are as follows:

Counter Type	Attributes	Example
Value Counter	ctr_xxxx: Complex object bundling all the counter values. ctrv_xxxx_current: Value of the current counter.	Counter Name: totalCPUTime ctr_totalCPUTime ctrv_totalCPUTime_current
Event Occurrence Counter	ctr_xxxx: Complex object bundling all the counter values. ctrv_xxxx_occurrences: Number of events since counter reset. ctrv_xxxx_persecond: Number of events per second in the moving average window.	Counter Name: NOMDocumentCreation ctr_NOMDocumentCreation ctrv_NOMDocumentCreation_occurrences ctrv_NOMDocumentCreation_persecond
Event Value Counter	ctr_xxxx: Complex object bundling all the counter values. ctrv_xxxx_average: Average value of the event in the moving average window. ctrv_xxxx_current: Last value of the last event. ctrv_xxxx_maximum: Maximum event value in the moving average window. ctrv_xxxx_minimum: Minimum event value in the moving average window. ctrv_xxxx_occurrences: Number of events since counter reset. ctrv_xxxx_persecond: Number of events per second in the moving average window. ctrv_xxxx_stddev: Standard deviation of the event values in the moving average window. ctrv_xxxx_total: Total event value since counter reset.	Counter Name: sendAndWaitTimer ctr_sendAndWaitTimer ctrv_sendAndWaitTimer_current ctrv_sendAndWaitTimer_occurrences ctrv_sendAndWaitTimer_persecond ctrv_sendAndWaitTimer_average ctrv_sendAndWaitTimer_minimum ctrv_sendAndWaitTimer_maximum ctrv_sendAndWaitTimer_stddev ctrv_sendAndWaitTimer_total

Operations

In Process Platform, getDefinedProblems and getOpenProblems are the default operations for all the managed components.

getDefinedProblems: Returns all the defined problems on the managed component in the form of a tabular data.
getOpenProblems: Returns all the open problems on the managed component in the form of a table.

Alerts

Alerts are designed to inform the administrator or the administration application about the major events in the application. For example, failure of database connection, starting and stopping of the service container are considered as alerts.

Alerts in Process Platform are exposed as JMX notifications, which can be monitored from any JMX management tool. In addition, these alerts are published through Log4J so that they can be forwarded as SMS or E-mail message. For more details, see Understanding Alert System.

Problems

A problem is defined as the state of the component, in which the functioning of the component does not happen in the expected manner. The defined problems can be viewed by invoking getDefinedProblems. The open problems can be viewed by invoking getOpenProblems from the operations. For more details, see Working With Problem Registry.

pp108 : Management Interface of Process Platform