1.3.1 Preparatory estimates and settings for running JP1/AJS3

Jobs

Total number of jobs per day

Estimate whether the system can handle the total number of jobs.

As a general rule, keep the number of jobs executed per day to no more than 10,000.

Even a well-designed system would be unable to process more than 50,000 to 100,000 jobs per day.

Reduce the number of jobs. We recommend no more than 10,000 jobs per day. Allow for reruns in case of job execution errors.

3.1.1 Job throughput

Number of jobs per hour at peak times

Estimate whether the system can handle the number of jobs at peak times. The limit is 5,000 per hour.

Reduce the number of jobs. We recommend no more than 500 to 1,000 jobs per hour. Allow for reruns in case of job execution errors.

3.1.1 Job throughput

Number of jobs when automatic retries are specified for jobs

Estimate the number of jobs, including the number of retry executions.

For example, if a maximum of five retries is set for a job, estimate the number of jobs as six (the original execution plus five retry executions).

Take the number specified for Maximum retry times into account. We recommend no more than 500 to 1,000 jobs per hour, and no more than 10,000 jobs per day.

3.1.1 Job throughput

Whether to use the standard output file, standard error output file, and file transfer

This affects the CPU load, memory usage, and job execution performance.

Try to keep the standard output file, standard error output file, and transferred files as small as possible.

7.10.1 Notes on the standard output file and standard error output file in the JP1/Automatic Job Management System 3 System Design (Work Tasks) Guide

Whether to append data to the standard output file and standard error output file as each job is executed

Appending data will increase the size of the files as each job is executed, causing problems such as increased CPU load when files are transferred.

If data is to be appended at each job execution, we recommend some measure such as periodically deleting the output files.

7.10.1 Notes on the standard output file and standard error output file in the JP1/Automatic Job Management System 3 System Design (Work Tasks) Guide

Maximum number of concurrently executable jobs

In Windows, be careful of jobs that take a long time to execute.

Due to the limitations of the desktop heap, the maximum number of executable jobs cannot be increased significantly in Windows. If jobs take a long time to execute and the maximum number is reached, you will need to take measures such as distributing the processing among multiple agents.

2.5.4 Maximum number of concurrently executable jobs, and

4.2.3 Changing the JP1/AJS3 service settings (Windows only)

Jobnets

Number of root jobnets executed per day

As a general rule, register no more than 4,000 jobnets for execution per day.

If the maximum is exceeded, allocate the root jobnets among multiple scheduler services.

4.3.2 Activating multiple scheduler services concurrently

Number of defined jobnets

Reduce the number of root jobnets.

When a large number of root jobnets contain only a few lower units, combine them into a single root jobnet wherever possible.

7.1 Notes on the number of root jobnets registered for execution in the JP1/Automatic Job Management System 3 System Design (Work Tasks) Guide

Job group hierarchies

As a general rule, keep the number of root jobnets in a job group to no more than 500.

Jobnet hierarchies

Create 2 or 3 levels of hierarchy, including the root jobnet, because more complex hierarchies result in reduced visibility.

Jobnet structure

Allow no more than 50-100 nested jobnets and jobs to be defined in one root jobnet or nested jobnet. Make sure that no more than approximately 500 nested jobnets and jobs are located below one root jobnet.

Number of logs to keep

Generations of execution results exceeding the set limit will be deleted when a root jobnet is executed. This affects the root jobnet's start performance.

Reduce the number of root jobnets that can be registered, bearing in mind that saved generations exceeding the limit will be deleted when the root jobnet is executed.

7.2 Relationship between number of logs to keep and performance in the JP1/Automatic Job Management System 3 System Design (Work Tasks) Guide

This affects the database size, response time, and resources.

Keep as few generations of data as your system will allow.

Jobnet and job operation

Continuous operation

If performed continuously, operations that place a high load on the scheduler service will further increase the system load, which could affect system operations.

Avoid continuous operations that entail jobnet or job execution. Pause for 2-3 seconds between each operation.

1.3.2 Operations that might affect JP1/AJS3 system operation, and

1.2 Notes on using commands in the manual JP1/Automatic Job Management System 3 Command Reference

Number of lower units

The amount of information to be processed increases in proportion to the number of lower units. This could affect system operations.

When the number of lower units exceeds 500, execute the jobnet when the work-load is low.

Wildcard characters in commands

If a command is executed with wildcard characters, the number of information items to process might increase, which might affect system operations.

Restrict the number of units targeted by the command before executing the command, such as by directly specifying the units that would be targeted by executing the command with wildcard characters.

Number of logs to keep

Operating on a jobnet that has many generations of execution results increases the amount of information to be processed. This could affect system operations.

If the target jobnet has a large number of generations, perform the operation when the work-load is low. Specify options to limit the time period and number of generations.

When canceling registration of a jobnet, you can also specify an option that deletes generations one by one.

Start conditions

Whether to use start conditions

Jobs with start conditions are used for transaction processing purposes and have a significant effect on the number of jobs being executed.

Accurately estimate the number of events issued and the peak traffic.

3.1.5 Event monitoring performance

Event jobs

Timing of monitoring JP1 event reception monitoring jobs

The status of the JP1 event-reception monitoring job must be Monitoring to receive JP1 events.

To link jobnets using a JP1 event, register the JP1 event-reception monitoring job at a set time before the JP1 event is sent.

7.6.1 Notes on the Receive JP1 Event job in the JP1/Automatic Job Management System 3 System Design (Work Tasks) Guide

Number of Monitoring Files jobs

This has a significant effect on CPU performance and the response time for Monitoring Files jobs.

When monitoring files over a network, the network performance is also affected.

Determine the file monitoring interval, using the equation in 3.1.6 Monitoring interval set when using the Monitoring Files job, and extend the monitoring interval accordingly.

3.1.6 Monitoring interval set when using the Monitoring Files job

Monitoring frequency

This has a significant effect on CPU performance and the response time for Monitoring Files jobs.

When monitoring files over a network, the network performance is also affected.

Extend the monitoring interval as appropriate.

Number of files monitored by Monitoring Files jobs

Timing of updates to files monitored by Monitoring Files jobs

Monitoring Files jobs cannot monitor files that are frequently opened and closed to append data.

Consider solutions such as creating and updating a separate file from the file that was last updated.

7.6.2 Notes on the Monitoring Files job in the JP1/Automatic Job Management System 3 System Design (Work Tasks) Guide

JP1/AJS3 - View

Automatic update

By default, displayed contents are updated automatically, and a database access request is generated on the manager host at every update.

The update interval is 300 seconds for the JP1/AJS3 - View window and Daily Schedule window, and 30 seconds for the Jobnet Monitor window. If performance is paramount, set a larger update interval or disable automatic update.

4.5.5 Setting the window display

Monthly Schedule window and Daily Schedule window

The number of display items and the amount of data transferred to the Monthly and Daily Schedule windows increases with the number of jobnets being monitored and the number of schedules.

Make sure that you complete the performance estimates in advance.

3.2.1 Estimating the data flow on the network

Number of connections

The load on JP1/AJS3 - Manager and on the network increases when more than 63 instances of JP1/AJS3 - View are connected.

Set a larger update interval or disable automatic update.

B. List of Limits