Benefits
The system can publish statistics at a granular level. Using a third-party monitoring application is beneficial, as you can perform a detailed analysis of the statistics.
| Current capability | Enhanced capability |
|---|---|
| The system currently consolidates the metrics reported by the JVMs. | The metrics from the JVMs are passed directly to the third-party application without consolidation. |
Summary of the metrics returned by the JVMs
is recorded in the YFS_STATISTICS_DETAIL table. |
Based on the properties that are configured
in the customer_overrides.properties file, the metrics
reported by the JVMs are recorded in the YFS_STATISTICS_DETAIL table
or passed to the third-party application, or both. To pass the statistics
to a third-party application, you can use the PLTStatisticsConsumer and IStatisticObject interfaces.
For more information, see the Javadoc. |
Consolidated metrics may not be useful for an
exhaustive analysis of the system performance. For example, if the scheduleOrder API
is monitored, you can find out the number of times the API was executed,
the minimum, maximum, and the average response times. However, with
the limited scope of analysis by using consolidated metrics, you may
not be able to determine the response time for 90th percentile of
the API calls, as there are no data points to calculate the required
information. |
Third-party applications can be used to plot
useful graphs with granular metrics. For example, if the scheduleOrder API
is invoked X times in a given time interval, the response time of
each invocation is passed to the third-party application, which can
later be used to plot the minimum, maximum and average response times,
or other metrics required by the user. This level of granularity provides
users greater control over the metrics, which can be readily used
to analyze the system performance. |