Performance Tuning for Content Manager

Chapter 11. Performance monitoring, analysis, and tracing 283

system performance and collecting a series of outputs in order to support the

diagnosis of a possible performance problem.

Sample workload statistics to collect include:

 Number of desktop and Web client users

 Number of each frequently performed operation (search, view, import,

doc-routing) per hour

 Average document size and number of pages

 Time duration of completing batch jobs such as load process

 Time duration of partial and full system backup

Server resource utilization statistics to collect include:

 CPU utilization

 Memory utilization and footprint

 Disk utilization and disk I/O utilization

 Network utilization

Collect data from various periods of the working day, week, or month when

performance is likely to be an issue. Some of the workload peaks might be:

 In the middle of the morning for users

 During late night for batch jobs

 During massive document loads

 Over the weekend during full system backup

Use measurements to collect data for each of these peaks in workload, because

a performance problem might appear during only one of these periods and not

during other times.

In the following sections, we discuss monitoring tools at operating system level

for each of the Content Manager base products (DB2, WebSphere Application

Server), and performance analysis and tracing techniques for Content Manager.

11.3 Monitoring tools at operating system level

Monitoring the OS from the perspective of tuning performance of a Content

Manager environment means to look at certain operating system measurement

counters that show use of OS resources. We specifically look at CPU utilization,

memory consumption, disk I/O, and network I/O.

Note: Always be aware that any monitoring activity has an impact on the

performance of the system being monitored.