Database Sizing

AgilePoint supports SQL and Oracle, has data stores, and we typically suggest you to employ your own corporate application database standards regarding initial size and security related to data store.

All information related to the audit trail of a process and all activities/tasks within are stored in the workflow database. The information related to the front-end side is NOT stored in the AgilePoint database. AgilePoint is able to save all information (audit trail) of completed processes into an archive database.

The recommended best practice for data storage is to develop your application's data model as you would normally (with full indexing, relationships, constraints, etc.) as a separate database from the workflow database. You would then store only the "workflow-related" portions of the data in the workflow database (as custom process attributes), so that the data can be used for such purposes as controlling the process flow, building email notifications, etc.

In general you can determine your data store capacity based on the following parameters:

Number of unique process templates
Average number of activity steps per process definition
Size of attachments
Average size of custom attributes per process (application data carried along the process)

The following database sizing formula can be used to estimate the space required for the AgilePoint Workflow Database.

Assumptions:

Assume a fixed-size process
Assume 50 steps are in the process

Database estimates:

Process Instance Table = 5016 bytes

Sum of above = 145 KB (As a general rule the size is generally < 60 KB)

Database size formula in KB: (145 + [workflow relevant data])*(number of processes)

Assumptions:

Assume average workflow relevant data = 1024 KB
Assume process instances = 100/day or 36500/year

Total database sizing estimate for a year: 1169*100*365 = 42668500 KB or 42 GB

The workflow relevant data is a big percentage of the estimate and varies depending on the application.

Workflow Relevant Data: [WF_CUSTOM_ATTRS; WF_LARGE_BINARIES; WF_LARGE_TEXT]

Typically you should monitor the above parameters in your dev/test environments and be able to use the above data for calculating database sizing requirements for your production environment.