Job scaling tab

On the Job scaling tab you can select to split the processing of Messages over several threads. This can be used to increase performance when processing large input batches containing several Messages.

Note:

The processing is done as separate jobs, and there is no way to merge the batch back automatically once it is processed. This means care must be taken on the output side so that files are not overwritten etc.

This job scaling works both when queues are used, and when queues are not used. Performance can be increased more when queues are not used compared to when queues are used. In some cases when using queues and job scaling with several threads for small Messages, performance could even decrease when enabling job scaling. This means tuning is necessary in each Project in order to avoid decrease in performance.

Setting	Description
Single thread	Use one single thread, and do not enable job scaling.
Variable trigger	Use a variable to determine how to split the input job. All Messages that have the same value for this variable will be processed in the same thread. There will be as many threads used as there are values of the variable. This can be used, for example, to split a batch of invoices by using the customer number as variable. All invoices that belong to customer A is processed in one thread, all invoices that belong to customer B is processed in another thread and so on.
Round robin	Use a fixed number of threads for the input job. Splitting is done in a round robin manner.

Script execution of Before and After job scripts

The following applies when you have enabled job scaling:

•	A script added Before the Runtime job is executed once per input job.

•	A script added After the Runtime job is executed once per thread. That is, as many times as the number of threads specified for Round robin or as the number of variable values assigned via the Variable trigger.