HP-UX Workload Manager User's Guide
Monitoring SLO compliance and WLM
Monitoring WLM with EMS
Chapter 10358
Metric status updates
EMS resources providing the status of WLM metrics include the
following. A resource instance exists for each metric, where met_name is
the name of the metric as specified in the WLM configuration file.
• /applications/wlm/metric_config/met_name/coll_argv
This identifies a specific metric’s data collector as specified with the
coll_argv keyword in the tune structure. For more information on
specifying the coll_argv keyword in the configuration file, see
“Specifying a data collector (optional)” on page 215.
• /applications/wlm/metric_config/met_name/cntl_smooth
This indicates a specific metric’s smoothing value as specified with
the cntl_smooth keyword in the tune structure. The resource value
can range from 0 to 0.999. A value of 0 (the default) signifies that a
running average is not taken for the data collected; in other words,
dips and spikes of data received by the data collector are not
smoothed out before WLM uses the data. The higher the value, the
more smoothing that is performed. For more information on
specifying the cntl_smooth keyword in the configuration file, see
“Smoothing metric values (optional)” on page 223.
• /applications/wlm/metric_config/met_name/cntl_avg
This indicates whether averaging is enabled for a specific metric, as
specified with the cntl_avg keyword in the tune structure. Possible
values are:
WM_METRIC_CNTL_AVG_ON Averaging is enabled
WM_METRIC_CNTL_AVG_OFF Averaging is disabled, meaning
that only the most recent
interval’s (wlm_interval) usage
data is used for a usage metric
For more information on specifying the cntl_avg keyword in the
configuration file, see “Controlling averaging in usage controllers
(optional)” on page 223.
• /applications/wlm/metric_status/met_name
This indicates the current value for a specific metric. In contrast to
the previous metric resources (which are constant for a given WLM
configuration), the metric_status value is dynamic, meaning that the
value changes periodically during a configuration’s deployment.