Concepts

The public JMX APIs are modeled by a single instance of ServiceDomainMBean, which has operations to check for monitored services and retrieve data from them.

A public set of POJOs provide additional objects and methods that, along with ServiceDomainMbean, provide a complete API for monitoring statistics.

The following sections provide brief descriptions of the POJOs and MBean. The Javadoc provides detailed descriptions. There is also a detailed description of statistics that are reported for resources.

Public POJO Objects

ResourceType

This object represents all types of resources that are enabled for service monitoring. There are four enum constants representing types: SERVICE, FLOW_COMPONENT, URI, and WEBSERVICE_OPERATION.

ServiceResourceStatistic

This object represents all business and proxy service resource types and the statistics associated with them. There are methods to get statistics for all resources or for a specified one.

ResourceStatistic

This object represents a resource for which statistics collection is supported. There are methods to get the name of the resource, the type, and the statistics.

StatisticValue

This object represents a statistic value for a resource. The monitoring system currently supports the following types of statistic values, both nested classes:

StatisticValue is an abstract class so that concrete objects representing count and interval statistic values can be derived from it. It includes getName() and getType() methods.

StatisticType

This object represents predefined types of statistics. There are three enum types: STATUS, COUNT, and INTERVAL.

ServiceDomainMBean

This is the only MBean exposed as part of the public JMX API. It provides methods to find monitored service and get and reset statistics.

Statistics Details

The following sections provide detailed information about statistics reported for each resource type.

Statistics details for Resource Type - SERVICE

A service is an inbound or outbound endpoint that is configured within AquaLogic Service Bus. It may have an associated WSDL, security settings, etc.

Table 2-1 SERVICE Statistics
Statistic Name	Type
message-count	count
error-count	count
failover-count	count
wss-error	count
response-time	interval
validation-errors	count
failure-rate	count
success-rate	count
sla-severity-warning	count
sla-severity-major	count
sla-severity-minor	count
sla-severity-normal	count
sla-severity-fatal	count
sla-severity-critical	count
sla-severity-all	count
pipeline-severity-warning	count
pipeline-severity-major	count
pipeline-severity-minor	count
pipeline-severity-normal	count
pipeline-severity-fatal	count
pipeline-severity-critical	count
pipeline-severity-all	count
throttling-time	interval
uri-offline-count	count

The statistics sla-severity-warning, sla-severity-major, sla-severity-minor, sla-severity-normal, sla-severity-fatal, sla-severity-critical, and sla-severity-all are collected for both proxy services and business services.

The statistics pipeline-severity-warning, pipeline-severity-major, pipeline-severity-minor, pipeline-severity-normal, pipeline-severity-fatal, pipeline-severity-critical, and pipeline-severity-all are collected only for proxy services.

Statistic “wss-error” provides Web Service security violations counts. It is applicable to both business and proxy services.
Statistic “validation errors” is only applicable to proxy service; it is not returned for a business service.
Statistic “failover-count” is only applicable to business service; it is not returned for a proxy service.
When the statistics of a managed server are retrieved from a cluster domain using the ServiceDomainMBean the statistics for proxy services will not contain sla-severity-normal, sla-severity-minor, sla-severity-major, sla-severity-warning, sla-severity-critical, sla-severity-fatal, sla-severity-all.

Statistics details for Resource Type– FLOW_COMPONENT

Statistics are collected for the following two types of components that can be present in the flow definition of a proxy service.

Pipelines are one-way processing paths consisting of stages that are executed sequentially against the current message. Stages are used to perform activities such as transformation, logging and publishing.

The pipeline-pair node ties together a single request and a single response pipeline into one top-level element.

A routing node consists of a set of routes. A route identifies a target service and includes some additional configuration options that determines how the message will be packaged and sent to that service. A routing node will result in at most one route being selected as part of request processing.

Table 2-2 FLOW_COMPONENT Statistics
Statistic Name	Type
elapse-time	interval
message-count	count
error-count	count

Statistics for pipeline and route nodes are returned as statistics for flow components. enum value ResourceType.FLOW_COMPONENT represents both pipeline and route nodes.
Thus there is no way for a client to check if the returned flow component is a pipeline or route node. The name of the flow component, however, may suggest the type.

Statistics details for Resource Type – WEBSERVICE_OPERATION

This resource type provides statistical information pertaining to WSDL operations. Statistics are reported for each defined operation.

Table 2-3 WEBSERVICE_OPERATION Statistics
Statistic Name	Type
elapsed-time	interval
message-count	count
error-count	count

Statistics details for Resource Type – URI

This resource type provides statistical information pertaining to endpoint URI for a business service. Statistics are reported for each defined Endpoint URI. The following statistics are reported.

Caveats

A client program will not know about newly added services that have monitoring turned on, or services modified to turn on monitoring, unless it periodically checks for such changes.
Reset operations should not be performed too frequently. BEA recommends that reset intervals be greater than 15 minutes.
BEA strongly discourages using this API in a concurrent manner with more than one thread or process. This is because a reset performed in one thread or process is not visible to another threads or processes. This caveat also applies to resets performed from the Monitoring Dashboard of the AquaLogic Service Bus Console, as such resets are not visible to this API.

Performance

Performance should be better than or equivalent to that observed in the Monitoring Dashboard of the AquaLogic Service Bus Console.

JMX Monitoring API Programming Guide