Show TOC

Procedure documentationMonitoring the Performance of Web Services Locate this document in the navigation structure

 

If critical business processes are based on Web services provided by ABAP systems, as a technical or application support manager, check whether the operations (Web service methods being called) are carried out with the required performance. You have the following options:

  • Getting an overview of the number of different Web service executions, and their distribution

  • Displaying average execution times in histograms: To get an indication of possible threshold values for SLA reporting, display the distribution of execution times of operations. For example, after having collected data for several months you see that most of the execution times are below 300 ms.

  • Getting indications of root causes of poorly performing operations: To identify the root cause of performance bottlenecks, display the average duration of processes hourly. You get an indication of whether the data base, the operation, the system, or the amount of data transferred is the cause.

Prerequisites

Procedure

Starting the Monitoring UI
  1. In the Technical Monitoring work center, choose Integration Monitoring.

  2. On the Interface Channel Monitoring tab page, select a scenario and choose the Web Services Monitoring pushbutton.

  3. Select the ABAP system.

    On the Overview tab page, all web services executed on the system are displayed for the selected time frame.

Displaying the distribution of Web services operations by usage

To identify the most frequently executed Web services and operations, use the Overview tab page.

  1. To display the number and percentage of executions of a Web service, position the mouse on the section of pie chart,

  2. To drill down, click on the section.

    The number and percentage of Web service operations are displayed.

  3. If configured: Click the operation type, to drill down to the users performing the Web service operations.

Displaying execution times of operations

To analyze the distribution of average execution times of operations, do the following:

  1. On the Metric Details tab page, choose Histogram.

  2. Optional: Filter on Web services and operations.

    Aggregate is the average of all operations.

  3. Optional: In the ICMon configuration, define an alert. To do so, configure the following metrics and define threshold values:

    • Number of WS executions below threshold

    • Percentage of WS executions below threshold: Calculates (based on the above metric) how many executions (in percent) of the configured web services in the last hour are below the configured threshold.

      For this metric you can configure, for each interface channel, the threshold in % when an alert is to be raised.

    Example Example

    The business unit requires that 94% of all executions of a Web service per hour are faster than 1500 ms.

    1. In the Configuration step do the following:

      1. Enable the Number of WS executions below threshold metric and set the threshold to 1500.

      2. Enable the Percentage of WS executions below threshold metric.

    2. Save your entries.

    3. In the Activation step, expand the channel and select the metric Percentage of WS executions below threshold.

    4. Set the threshold for the metric to 94%.

    For more information about accessing the guided procedure, see Configuring an Interface and Connection Monitoring Scenario.

    End of the example.
Getting indications of root causes of poorly performing operations

To analyze the distribution of execution times of operations, do the following:

  1. On the Metric Details tab page, choose Metrics.

    The following work process parameters are displayed (average time in milliseconds):

    • DB_TIME: Time to write data to the database

    • RESPTIME: Response time, actual processing time of the operation

    • WAIT_TIME: Time a work process has to wait before the system provides a process. The wait time is an indication of the overall system performance.

    • MESSAGE_SIZE: Amount of payload data (in bytes) transferred when the operation is executed

  2. Filter on Web services and operations.

    Aggregate is the average of all operations.

  3. Optional: Adjust the time frame, or choose Configuration (on the right side of the screen), to switch the 24h profile on or off.