Performance Charts
It is therefore a good idea to regularly monitor a number of performance
Instrument | Description |
|
|
Process Instances | number of active process instances |
Activities | number of activities in the ready or active state |
Average response | the average time it takes for the engine to complete a request |
time | made by a session (such as OpenSession, StartActivity, and |
| CompleteActivity) |
Average transaction | the time it takes the engine to commit a change of state (complete |
commit time | a transaction involving the engine database) |
Allocated memory | the amount of the active memory currently allocated to objects |
| created by the primary engine unit |
Peak allocated | the amount of allocated memory remaining after the most recent |
memory | memory reclamation (probably the best measure of active |
| memory utilization) |
|
|
These instruments indicate how performance might be impacted by an increasing load on the engine.
For example, as the number of sessions increases over time, you would expect the average response time to increase. At some point, you might decide to add an additional engine to your workflow system to help balance the load.
As the number of process instances and activities increase over time, you would expect the average transaction commit time to increase. In this case you might decide to start up additional database services to speed throughput to the engine database.
In many situations these factors have complicated interactions and you have to monitor these indicators carefully under varying conditions, analyzing the source of any performance bottlenecks.
Viewing Performance Indicators
You can view performance indicators for any running engine in the Performance window.
