GiAPA

GiAPA from iPerformance is able to identify performance problems in the iSeries and to demonstrate where improvements are necessary. Do you experience problems at specific times of the day?
 
Many iSeries users experience problems at specific times during the day. This is why these users wish to find out why e.g. response times are enhanced and why certain batch runs take so long.
 
The issue is not necessarily poor coding
Even though the code of the company's iSeries applications is OK, significant performance enhancements can be obtained via GiAPA. The underlying cause is often that a traditional code has been applied rather than methods according to which performance is the primary target.
 
SOSY is able to generate reports on the performance data that supply an overview of the general system status at the specific point of time. Contact us if you e.g. need an annual status report.
 
Results obtained by GiAPA

• CPU response times reduced from minutes to seconds
• Batch runs reduced from hours to minutes
• Forecast on capacity utilisation as the number of users increases
• Postponements of expensive hardware upgradings

- GiAPA version 2.0 with a host of exiting new features

The newly released GiAPA version 2.0 from iPerformance has been improved and upgraded in several areas. The most important new feature is that we can now produce performance measurements of Java applications running on System i at the thread level, a major requirement from many of our users.

Thread level information
The many threads in Java jobs have previously been a major challenge in regard to performance measurements in Java programs. Starting with OS/400 version 5.4 it is possible to collect and analyse thread level information for Java jobs. GiAPA 2.0 has been enhanced to take advantage of this. Define critical levels for CPU% or disk I/O and GiAPA will identify which parts of your Java applications exceed these values. GiAPA supplies information on the Java class names, method tupes, and method name.

GiAPA basics

It is important to highlight the fact that GiAPA 2.0 still only uses 0,1% CPU, even though data on all applications are collected every 15 seconds. Whenever a performance problem is found, based on the defined threshold values for e.g. CPU% or disk I/O, GiAPA creates a "hotspot". Multi-fold performance analyses can then be made dynamically based on the identified hotspots.

Info about temporary indices
GiAPA 2.0 also automatically collects data when a temporary index is created. This information can be used to evaluate whether some of the temporary indices should be permanently defined, so that they do not need to re-create them each time they are run.

40 user parameters
Version 2.0 gives a much increased ability to set up parameters. Version 1.0 had fifteen user parameters – version 2.0 has forty. The resource requirement for a job can be defined before hotspot data is collected, also the number of threads for which data is to be collected, as well as the required number of call stack levels.

Coldspots
The concept of coldspots is introduced in GiAPA version 2. Coldspots pinpoint jobs or transactions which do not complete within 30 seconds, but which keep on running, using very few resources. This could e.g. indicate that a job is locked.

Other improvements...
Apart from these four important changes, GiAPA also offers the following improvements:

·        Complete call stacks information for up to 99 levels

·        Support for up to 99999 open databases and 4 billion records

·        Program and procedure names for up to 255 characters

·        Information on usernames for jobs which are run by multiple user profiles

·        Paging statistics per job (allocation/deallocation of storage)

Get the GiAPA poduct leaf here

See more at the iPerformance site here