This
test monitors the processing of requests by a JRun server. The measures
this test generates are discussed below:
| Measurement |
Description |
Measurement
Unit |
Interpretation |
| Avg_queue_time
|
This value indicates the average time a request spends waiting for
processing by the server. |
Secs
|
A low value here would indicate that the server is performing
optimally. An increase in queuing delay reflects a server bottleneck.
For a JRun server 3.0, consider increasing the jcp.endpoint.main.active.threads
property in the local.properties file to ensure that additional
threads are available to process incoming requests.
In case of a JRun server 4.0, modify the activeHandlerThreads
attribute of the WebService service in the jrun.xml file
located in the <JRUN_HOME_DIR>/servers/<SERVER_NAME>/SERVER-INF
directory.
A number of
compute intensive tasks can also end up increasing the queuing time
for a request.
|
| Avg_proc_time
|
This value indicates the average time taken by the server to process a
request. |
Secs
|
A high value here would indicate that processing time per request is
very high, as the requests might be performing complex/time-consuming
computation. Consider determining which of the server side processing
tasks is compute intensive and explore ways of minimizing the server
processing overhead.
|
| Avg_response_time
|
This indicates the average time spent by a request in queuing and
processing. |
Secs
|
An increase in response time can occur because there are too many
simultaneous requests or because of a bottleneck with any of the
applications executing on the server.
|
| Data_read_rate
|
This indicates the total rate of data received by the server for all
incoming requests. |
KB/Sec
|
An unusually high value indicates an increase in workload to the
server. The Avg_queue_time metric indicates whether the increased
incoming data rate is impacting the server performance.
|
| Data_write_rate
|
This indicates the total rate of data sent by the server in response
to all incoming requests. |
KB/Sec
|
An unusually high value indicates an increase in workload to the
server. The Avg_queue_time metric indicates whether the increased
incoming data rate is impacting the server performance.
|
| Delayed_request_count
|
This value indicates the number of requests delayed at the
server.
|
Number
|
While the Avg_queue_time gives an indicator of how long requests are
being queued, the Delayed_request_count provides an indicator of how
many requests were queued at the server over the last measurement
period. This value should be close to 0 for peak performance.
|
| Dropped_request_count
|
This indicates the number of requests that were dropped since the
server could not process them or queue them. |
Number
|
This value should be close to 0 at most times. . Therefore, if a JRun server 3.0 drops a large
number of requests, consider increasing the parameter jcp.endpoint.main.max.threads,
in the local.properties file.
In case of a JRun server 4.0, modify the activeHandlerThreads
attribute of the WebService service in the jrun.xml file
located in the <JRUN_HOME_DIR>/servers/<SERVER_NAME>/SERVER-INF
directory.
This parameter indicates the
maximum number of requests that can be processed or queued at any
instant.
|
| Request_rate
|
This indicates the rate of requests handled by the JRun server |
Reqs/sec
|
This vailue is an indicator of server workload across all its
connectors.
|
| Note:
If a new server instance is added to a JRun application server
while its being monitored, it will take at least an hour for the new
instance to appear in the eG monitor console. |