1.1.2 Collect performance data according to its characteristics
Performance data is collected in the form of a record. Each record is further divided into units called fields. Records and fields are generically referred to as data models.
According to their characteristics, records can be classified into three types. The kind of performance data to be collected by each type of record is defined by PFM - Agent for Platform. Using PFM - Web Console, the user can select the kinds of performance data to be collected as records.
The following record types are available in PFM - Agent for Platform:
-
Product Interval record type (hereafter referred to as PI record type)
Performance data for a specific duration (interval), such as the number of active processes during every minute, is collected as a PI record. Use the PI record type when you want to analyze the changes and trends in the system status over time.
-
Product Detail record type (hereafter referred to as PD record type)
Performance data that indicates the system status at a given point in time, such as detailed information on the processes that are currently active, is collected as a PD record. Use the PD record type when you want to determine the system status at a given point in time.
-
Product Log record type (hereafter referred to as PL record type)
Log information from applications being run on UNIX or databases is collected as a PL record.
For details about these record types, see 9. Records.