There are common assumptions people make when they look
at Process Monitor data in the Process Admin Console that might lead
them to interpret their process performance incorrectly. Learn about
the areas of Process Monitor where users are most likely to misinterpret
what the data represents.
Completed Steps column
The completed steps
column represents the current execution state of the process or service
step that is received from the process engine. It means only that
the process or service engine has completed execution of the step,
but it does not mean the step itself is actually completed, as the
step might be waiting for a response from a user or the system.
Process App column
The Process App column in the Active services table lists
the process app or toolkit that contains the service at design time (in the Workflow Center); it is not the
runtime process application. For example, in the following image, although the service Default Human
Service is a task of the BPD instance at run time, it is shown as part of
UI(8.6.6.0) because it is in the UI(8.6.6.0) toolkit.
Total Steps count
The total steps that are
shown in the Process Monitor do not correspond to the total number
of steps as defined in the business process definition. The total
steps count includes the number of times the service is generated,
even when the user does not actually run the service. For example,
if a IBM® Heritage Process
Portal(deprecated) user starts a
task that is a human service, the corresponding coach page appears.
If the user closes the coach page immediately without clicking anything
else, the step total increases by one.
Halt process and halt service buttons
It is
typically difficult to see the halt process and halt service buttons,
as they appear only when the service is actually being executed by
the engine. When you see these buttons, the process or service is
typically running in a loop, or the service is taking a long time
to run.
Inconsistent monitoring data
In certain
situations, you might see data that seems contradictory. For example,
the total number of currently executing processes or services might
be zero, but there are running processes or services. Figure 1. Processes
page
Figure 2. Services
page
Figure 3. Summary page
For example, the Processes page
shows no active processes (Figure 1), but there are active processes
in Heritage Process Portal,
and the Services and Summary pages (Figures 2 and 3) show active process
apps in the table.
These results might occur in the following
scenario:
The system is not monitoring when the process instance is created.
The system starts monitoring.
The task in the process instance starts running.
A Process Monitor record is created when the event occurs. The
process instance record is created when the instance is created. In
the preceding scenario, when the process gets created, the corresponding
record is not created, since the system is not monitoring. But the
service event occurred after the system started monitoring. So there
are service records but no process record. This situation might also
occur when the server restarts.
Exit time calculation
If a process or service
is still executing when it is exported, there is no exit time, and
the system uses the current time as the exit time to calculate the
duration. As you refresh the information on screen or through the
JMX api, the duration time is updated. This explains why when a process
or service that is executing is exported more than once, the duration
time that shown in the exported file can vary.