The default pmtrace behavior is to provide point trace data to the trace PMDA, using the tag argument as the identifying name associated with each trace point. The tag then becomes an instance identifier within the set of trace.point metrics.
The -e option allows an arbitrary command to be executed. This command will be measured as a transaction since it has well defined start and end points. The information is made available through the trace.transact metrics.
Trace data can be sent to the trace PMDA running on host , rather than the localhost, using the -h option. This overrides use of the environment variable PCP_TRACE_HOST .
The -q option suppresses messages from a successful trace, so that pmtrace runs quietly.
The -c option allows an arbitrary counter value to be exported through the trace.count metrics, while the -v option allows an arbitrary floating point value to be exported through the trace.observe metrics
The -S option enables internal debugging and tracing. The value of state is a bit-wise combination of debug flags as defined in pmtracestate (3), and may be specified using the decimal or hexadecimal syntax prescribed by strtol (3).
The pmtrace exit status is always zero except when the -e option is in use, in which case the exit status of command is returned.