Command line batch jobs in EcoSys have slightly different considerations for performance tuning than the web application does.
Java heap size
The maximum Java heap size for a command line batch job can be just enough to support that job, typically much smaller than required for the web application. The minimum heap size for the command line job may also be much less than the maximum, as Java allocates only the memory it requires as it runs.
Database connection pool
The database connection pool for command line batch jobs is configured separately from the web application. Parameters for the command line’s database connection pool are specified in the FMServerSettings.properties file. Typically, the pool consists of 5-10 connections unless heavy parallelization is used.
Parallel Load/Save (PULS)
By default, the PULS multi-threading feature is not enabled for command line batch jobs. You may enable this feature as needed, based on the specific requirements for each job. Often, command line batch mode is better suited to running more slowly, and using fewer system resources. As such, parallelization may not be appropriate.
Performance metrics
Server metrics are not captured during command line batch jobs because there is only one job executing in the Java VM for the life of that JVM. You can monitor resource utilization at the application server (operating system) level instead.