An nng_stat represents a statistic.
All statistics have names (nng_stat_name()) and
descriptions (nng_stat_desc()), and are
typed (nng_stat_type()).
Most statistics are numeric,
and thus carry a value (nng_stat_value())
and frequently also a unit that the value measures (nng_stat_unit()).
Some statistics however, are simply strings (nng_stat_string()),
and thus carry no numeric value.
Statistics are organized as a tree, and any given statistic can have siblings
(nng_stat_next()).
Note however that generally only NNG_STAT_SCOPE statistics, which are
act as placeholders in the tree (and carry no value),
will have children (nng_stat_child()).
A tree of statistics is collected using nng_stats_get(),
and can be freed when no longer needed with nng_stats_free().
This collection process is generally performed in a way to minimize impact
to running operations, but there is still some impact caused by collection
of statistics.
The time when a statistic’s value is captured can be obtained using
nng_stat_timestap(), which is useful for
measuring rates of change in certain statistics.
|
Note
|
The presence, names, and meanings of any given statistic are subject to change at any time. These statistics are provided as an aid for debugging, and should generally not be relied upon for programmatic behaviors. |
|
Note
|
Statistics may be disabled by build-time configuration options, in order to reduce program size and run-time overheads. |