Predicting AgilePoint Server Resource Usage

Although the AgilePoint Server application normally runs constantly and requires a certain amount of server resources, it is optimized so that server resources are not normally consumed for data and processes that are not currently being actively worked on, interacted with, or updated. This allows AgilePoint Server to potentially interact with millions of separate process instances over time using limited CPU and RAM at any given time. However, when trying to estimate usage patterns and predict server load and resource usage, the following points should be considered.

AgilePoint Server's core engine is normally active and running at all times, even if no users or external systems are currently interacting with AgilePoint. This engine is responsible for various functionality, such as handling requests to the Workflow API and for scheduling and coordinating time-dependent events (e.g. reminder emails). The core engine is optimized to consume as little of the server's resources as possible, but it will require some portion of the server's CPU cycles and system memory.

However, in addition to the minimum server resources used by the core engine itself, additional server resources will be required and used when specific work or processing occurs. Examples of things that will initiate additional processing and (temporarily) increase the server's load include:

  • Processing of requests resulting from direct interaction with the AgilePoint system by users.

    For more information, refer to Memory Utilization.

  • Processing of requests resulting from direct interaction with the AgilePoint system by external systems
  • Processing (if necessary) of the escalation scenario defined by the process model for a task or activity that has "timed out" (i.e. become overdue)
  • Processing (as needed) of time-dependent functionality (e.g. reaching the end of the configured time interval for a Delay AgileShape)
  • Processing (as needed) of scheduled and/or recurring operations, such as Active Directory synchronization, or archiving of old workflow data
  • The escalation scenario defined by the process model for a task or activity that has "timed out" (i.e. become overdue)

All of the above represent a request (whether direct or indirect) for AgilePoint to perform additional work and processing, and the amount of server resources required and used to perform that processing will depend on the exact nature of the processing that needs to be performed.