Use the Service Monitor widget to measure response time and throughput over a defined period
of time for the services exposed and invoked by your module.
The
Service Monitor widget can display graphs
for both response time and throughput so you can compare and correlate
the data. It also displays a table with operation statistics. Both
graphs plot over a defined time window and are regularly updated according
to the specified refresh rate. As new data arrives, the oldest data
is removed from the graph.
- Response Time graph
- Graphs the response time for calls in the selected operation or
operations so you can determine how long the service takes and whether
its duration degrades over time. The Y-axis indicates the response
time, measured in milliseconds. The X-axis indicates the elapsed time
since monitoring began (for example, the window of time
you are monitoring), measured in minutes or hours. You can plot the
minimum, maximum, or mean response times for the specified time unit.
For service operations with a two-way asynchronous service implementation,
the response time indicates only the time the operation takes to handle
the re quest. It does not measure the total time elapsed before a
response is sent.
- Throughput graph
- Graphs the throughput for calls in the selected operation or operations
so you can see how often a service is called and whether expected
throughput benchmarks are being met. This graph indicates the number
of calls completed over a specific period of time (measured in seconds
or minutes). The Y-axis shows the number of calls completed per time
unit, with the time unit measured in seconds or minutes. The X-axis
shows the elapsed time, measured in minutes or hours.
- Statistic Measurements Table
- If you select Show statistics in the widget configuration options, this table provides operation
statistics cumulated over the last second or minute and over the duration
of service monitoring. The cumulated statistics include the response
time, throughput count, and failure count for each monitored operation.
Tips for using the graphs
The following
tips can help you use the Response Time and Throughput graphs more
effectively.
- Customize the time period and refresh rate for the graphs. You can adjust the refresh rate to increase or decrease the frequency
of updates to the graphs; remember that a faster refresh rate can
affect performance. You can also specify the length of time for which
you want to plot data on the graphs; graph time length is measured
in either minutes or hours.
- Restrict the upper limit for the Y-axis. The Y-axis is
scalable to optimize graph details; it dynamically adjusts to show
the minimum and maximum values since monitoring began. In addition,
you can configure an upper limit to restrict the Y-axis further in
situations where you have a few response time or throughput values
that greatly exceed the rest. In this situation, the graph does not
display any measurements greater than the upper limit.
- See exact measurements for a point on a graph. Hover over
any point in the Response Time or Throughput graph to see a window
with the operation name and specific response time or throughput data.
- Set a threshold for response time and throughput. For each
operation you monitor, specify a threshold for the response time,
throughput, or both. Calls that exceed the threshold are visually
distinguished on the graphs.
- Switch between different graph styles. The Service Monitor
widget offers three different graph types for viewing response time
and throughput data: Line (the default), Column, and Spike. You can
change the graph type for each graph dynamically from the drop-down
menu on top the graph.
- Monitor multiple operations on one graph. You can use the
graphs to monitor multiple operations at once (for example, to compare
the response times for a set of related operations). The Response
Time and Throughput graphs can each plot up to five operations.