In properly constructed composites, all service entry points to the different components such as BPEL, BPMN, Business Rules, and so on, are fronted by a mediator. The mediator component of Oracle SOA Suite 12c is intended to serve as an intracomposite router that handles interactions between its components. It therefore has the potential to become a constraint in infrastructure setups that involve a high volume of message processing. Fortunately, the Mediator engine dashboard provides a snapshot of processing statistics for various metrics such as count and execution times for one-way invokes, transformations, validations, condition evaluations, and so on. The Mediator engine dashboard also shows all the mediator components deployed to the infrastructure along with their statuses. The dashboard can be viewed by navigating to soa-infra | SOA Infrastructure | Service Engines | Mediator.
As with the BPEL engine, the performance of the Mediator engine and the components deployed to it depends on the configuration of engine properties. There are five groups of properties available to configure within the Mediator service engine:
These properties can be accessed by navigating to soa-infra | SOA Administration | Mediator Properties. The following table describes each of the configurable properties available:
Property |
Category |
Description |
---|---|---|
Audit Level |
Logging |
Setting this property to |
Metrics Level |
Logging |
This property determines whether Dynamic Monitoring Service (DMS) metrics should be collected for Mediator services. DMS metrics are used to measure the performance of application components. See The DMS Spy Servlet in Chapter 6, Monitoring Oracle SOA Suite 12c, for more details. |
Parallel Worker Threads |
Mediator Engine |
This property sets the number of outbound threads for parallel processing. This does not impact sequential services. |
Parallel Maximum Rows Retrieved |
Mediator Engine |
This specifies the number of rows to retrieve per iteration for parallel processing. Oracle documentation recommends setting this value to 50 to 100 times that of the |
Parallel Locker Thread Sleep |
Mediator Engine |
This specifies the idle time (in seconds) between two successive iterations for retrieving rows when there is no message for parallel processing. We almost always recommend setting this to |
Error Locker Thread Sleep |
Mediator Engine |
This is similar in concept to |
Parameters |
Custom |
This allows you to specify custom configuration properties. For example, in resequenced messages, it is possible to configure the buffer window for the time window in best-effort resequencing by adding the |
Container ID Refresh Time |
Health Check |
This is the interval (in seconds) in which the heartbeat thread checks the status of the Mediator Service Engine and announces its presence to other servers in the cluster. This is internally accomplished by updating the timestamp of the unique identifier maintained in each Mediator Service Engine. The default setting is 60 seconds. |
Container ID Lease Timeout |
Health Check |
This is the interval (in seconds) in which the heartbeat thread checks if there are other unique identifiers that have not been updated. |
Resequencer Locker Thread Sleep |
Resequencing |
This specifies the sleep time (in seconds) for a deferred locker when there is no message in the database. |
Resequencer Maximum Groups Locked |
Resequencing |
This specifies the maximum number of group rows retrieved for each locking cycle. |
Resequencer Worker Threads |
Resequencing |
This specifies the number of resequencer threads. |
3.17.174.0