When the WebStatistica Service is running, it will start up sub-processes of Statistica to service requests to run analyses and scripting requests. These can be seen as statcf.exe in the task manager.
A request to run an analysis or user interface script is called a job. The two groupings of Statistica instances are the “long-running” and “short-running” job queues:
1. The short-running queue is used for scripts that will take a short period of time. Examples include:
- Tasks such as generating a user interface, such as the list of files the user sees when first logging on
- A user selecting parameters for an analysis
2. The long-running queue is used for jobs that might take a longer time to run, including most analyses. Batch jobs will always run on the long-running job queue.
Separating requests into the two queues ensures that user-interface operations, such as logging on or selecting analyses, will respond more quickly when a backlog of analytical jobs is running.
NOTE: A short-running job can use a long-running instance if the short-running queue is full.
You can control the minimum and maximum number of Statistica instances in both the long running and short running queues.
The minimum value specifies the number of Statistica instances that will always be kept running even if no requests are pending. Keeping these instances running makes the system more responsive because it does not have to load an instance of Statistica to service the request.
The maximum value controls the maximum number of Statistica instances that can be started when many requests come in. When the maximum number of instances of Statistica is in use, new requests will be queued and serviced in the order they are received. The exception is that a batch job will always be a lower priority than an interactive job. Once no pending requests have been received for a period of time, the number of Statistica instances will go back to the minimum level.
In most cases, you should set the minimum and maximum number to the same value within the same job queue to ensure the fastest response and the most predictable resource usage. However, having a maximum number that differs from the minimum enables the system to dynamically adjust the number of Statistica instances when the load gets heavier.