YADE JITL job should be capable to process orders without exhaustion of resources

Current Situation
An API job in a job chain executes orders without a break. Therefore the task of the job never ends. After an unspecified number of orders you get the error:

COM-80020009  DISP_E_EXCEPTION [step] [] [Z-JAVA-105  Java exception java.lang.OutOfMemoryError("unable to create new native thread"), method=spooler_process()Z [], in Remote_module_instance_server::step], method=call

How to to reproduce the problem
Copy the attached file (force_new_task.js) to the folder where the API job is stored and configure the job with the following monitor script and a parameter "max_num_of_orders_per_task":

<job  order="yes" stop_on_error="no">
    <params >
        <param  name="max_num_of_orders_per_task" value="10"/>
    </params>

    <script .../>

    <monitor  name="forceNewTask" ordering="1">
        <script  language="javax.script:rhino">
            <include  live_file="force_new_task.js"/>
        </script>
    </monitor>

    <run_time />
</job>

This script ends the current task after the max. number of orders which is set by the above parameter.
The next order creates a new task.

Desired Behavior
The YADE JITL job should be capable to process orders without exhaustion of resources (threads, memory etc.).

Resolution

• JobScheduler Version 1.8
• One reason could also be SSH server settings where some SSH connection are dropped after pre defined time limit i.e. every 12 hours