SMF 100 machine learning data

The IBM® Z Operations Analytics machine learning system uses SMF 100 data to provide insight into normal and abnormal resource usage in Db2® for z/OS® subsystems, based on the CPU time, distributed data facility (DDF), locks, latches, and storage usage.

Data stream definition for SMF 100 machine learning data

Tip: These data streams can be defined only for the IBM Z Operations Analytics machine learning system.

For prerequisite requirements for defining SMF data streams, see Preparing the Configuration Tool to support SMF record types for Z Operations Analytics.

Table 1 indicates the configuration values to use in defining these data streams in the IBM Z Common Data Provider Configuration Tool.

Table 1. Data stream definition for SMF 100 machine learning data
Type of node in the policy Required configuration value
Data Stream All of the following values:
  • A_DB2_DB_SYS_I
  • A_DB2_DB_BIND_I
  • A_DB2_DB_I
  • A_DB2_BP_I
  • A_DB2_SHR_LOCK_I
  • A_DB2_SHR_BP_I
  • A_DB2_STORAGE_I
  • A_DB2_LATCH_I
To select these data streams in the Configuration Tool: In the Select data stream window, click IBM Z Operations Analytics – Machine Learning > Database > Db2, and select the check box for each stream.
Filter Transform Not required
Subscriber Machine learning system only.

For the appropriate values, see Subscribers for each type of source data.