All measured values are displayed in the report, even when they do not measure the performance of the monitored system. Therefore it is possible to use custom timers and counters that do not influence the performance rate. You can however view such values in reports and thereby gain additional data about the state of a system.
Page timers influence the performance rate either by comparison to previously measured data of the same timer or by comparing the timer value to its boundary, as defined in the monitor settings. See Calculating Health for details.
In general, transaction response time is also used for performance rate calculation. To gain comparable values, think times are not counted toward transaction response time. Time spent in wait statements is also ignored.
If transaction response time is not to influence performance rate set in the client monitor settings to "display only" or "no results". In the first case, the response time will still appear in the report. This can be useful for transactions that query performance data rather than simulate business transactions. For example, the time it takes to read performance values from a statistics page, or the time that is required to query performance counters using the PDCE shouldn't be reflected in the performance health of a monitored system.
Custom timers can also be used for performance rate calculation. However by default they are shown in the report and do not influence the performance rate. If they are to influence performance, their name must begin with Sv_. The name in the report shows the prefix to indicate which of the custom timers influence the performance rate and which do not.
HTIMER hMyTimer
...
hMyTimer = TimerCreate("Sv_LDAP Ping@"+sHost);
TimerStart(hMyTimer);
if WebLdapConnect(hLdap, sHost, nPort, nFlag) then
WebLdapBind(hLdap, NULL, NULL);
WebLdapDisconnect(hLdap);
end;
TimerStop(hMyTimer)
TimerDestroy(hMyTimer)
Custom counters can be used to indicate performance, accuracy, or availability of a system. By default, they are only displayed in reports and do not influence health rate. When a custom counter measures performance data, it must begin with the prefix Pe_. In such cases, boundaries should indicate whether large values are considered good or bad.
MeasureSetBound("Pe_CPU", MEASURE_COUNTER_CUSTOMCOUNTER, 1, float(nBoundCpu1));
MeasureSetBound("Pe_CPU", MEASURE_COUNTER_CUSTOMCOUNTER, 2, float(nBoundCpu2));
MeasureIncFloat("Pe_CPU", fValue, "%");
For custom counters to influence accuracy, the prefix Ac_ is used. The prefix Av_ creates a custom counter that influences the availability rate. When such a counter is set, the monitor reports a problem in the accuracy or availability of the monitored site.
MeasureInc ("Av_Host not reachable");