Hitachi

JP1 Version 12 JP1/Performance Management Planning and Configuration Guide


3.2.2 Flow of performance data collection

Organization of this subsection

(1) Flow of processing in performance data collection

Figure 3‒10: Flow of processing during performance data collection

[Figure]

Flow of processing:

  1. The performance data is collected.

    The Agent Collector service or Remote Monitor Collector service collects the performance data and manages the data in record format.

  2. The performance data is stored.

    (a) For real-time data

    The Store database does not store the collected performance but the data appears in the real-time reports.

    (b) For historical data

    The Agent Store service and Remote Monitor Store service store the collected performance data in the Store database. The system uses this data to display the historical reports.

(2) Life cycle of performance data

PFM - Agent and PFM - RM collect performance data at the following times:

PFM - Agent or PFM - RM might determine that the monitored target of the performance data collected at this time is the same target of the performance data previously collected. In this case PFM - Agent or PFM - RM treats the data as if they are the same records in the same field even if the collected data is from different targets.

Example regarding consistency:

In the case of the Process Detail (PD) record of PFM - Agent for Platform (Windows), the system uses the process name and process ID to judge whether performance data is from the same monitored target.

When the process name and process ID are the same as those of the data collected previously, PFM - Agent for Platform (Windows) recognizes the performance data collected at this time as data of the same process. Even when the process disappears during a collection interval and is later regenerated, if the process name and the process ID are the same as those of the data previously collected, PFM - Agent or PFM - RM recognize the data as the performance data of the same process. In such cases, the performance data becomes inconsistent.

Figure 3‒11: Example of performance data consistency

[Figure]

Point:

The period during which performance data maintains consistency is called the lifetime of the performance data. Take into account the lifetime of performance data when specifying the refresh intervals for real-time data and collection intervals for historical data.