Agent server level segregation

In the agent criteria approach for segregating workload, you can set the filter_criteria values against the Task Queue Filter Criteria parameter to create multiple criteria as required, associate them to same or separate servers, and then run one or more instances based on the throughput requirements of the servers.

Agent server level segregation is an alternative option to segregate the workload. In this approach filter criteria is set as a server startup parameter.

The agent services that are associated with the agent server might or might not have configured Task Queue Filter Criteria among the criteria parameters. The agent server level filter criteria supersedes any Task Queue Filter Criteria that is defined among the agent services. The advantage here is that you can specify the filter criteria while starting the server without changing any configuration as many instances of the agent server can be started based on the segregation requirements and to meet the throughput requirements.

The agent server startup script accepts a filtercriteria parameter to specify one or more comma-separated segregation patterns. You can pass the parameters as shown here:

As a JVM argument

agentserver.sh -jvmargs "-Dfiltercriteria=<segregation patterns>" <agent server>

As an agent server argument

agentserver.sh <agent server> -filtercriteria <segregation patterns>

All task queue agents that run on agent server matches FILTER_CRITERIA=<filtercriteria value> to get the jobs for processing.

In the agent criteria behavior, VOID is also a valid filter criteria. If you have multiple segregation patterns to be processed, you can use them as comma-separated values.

agentserver.sh <agent server> -filtercriteria "VeryLarge,VOID"