To access information in the statistics store, you need to use a statistics store identifier. Statistics store identifiers name system resources, statistics, and events. These identifiers can also specify operations on statistics and formatting of events. These identifiers are used at the command line, in the API, in the sheet creation dialogs of the System Web Interface, and in metadata files.
This appendix describes how to use statistics store identifiers to represent system resources, statistics, events, and topology.
Many identifiers are defined by default and added automatically when associated resources are added such as a disk or other device. Other identifiers are added when you run an application because the application uses statistics store APIs. You can add new identifiers to the statistics store and specify how data will be collected for them as described in Adding Custom Data to the Oracle Solaris 11.4 StatsStore and System Web Interface.
A statistics store identifier is also called an SSID. An SSID has different formats for statistics, events, and collections. All SSIDs specify a class. SSIDs can optionally specify a system.
Each component of an SSID is preceded by the three-character //: separator. While names in an SSID can include / and :, the three-character //: sequence is reserved.
Each component of an SSID consists of a keyword, followed by a period (.), followed by a name that identifies one or more instances of that component. Each component can have its own special characters or namespace restrictions for instance names. To represent multiple SSIDs in one SSID, you can use slice and wildcard notation. The * character can appear anywhere and matches to the next //: separator. See Displaying Resource and Statistic Slices for information about the s (slice) keyword.
The system component distinguishes statistics from different systems. The system-name is the name of the system that produced this statistic. Some classes have no resources on a particular system. An example of a system-name is example.com/localhost, where example.com is the name of the system. When no system component is specified, the local system is assumed.
A class is a type of object such as a device, a kernel statistic, or a service. A class-name represents all statistics for objects of that type. These statistics can be optional or required for that class. Some classes have no required statistics.
A class can specify that it is a subclass of other classes and inherit the namespace and statistics of the parents.
Examples of class-name values include app, cpu, dev, io, link, pg, pset, svc, system, and zone. Use the following command to list all the statistics store classes on the system:
$ sstore list //:class.*
For more information about the sstore list command, see Using Command Line Interfaces.