|
1. Data Collection.
GiAPA's performance data
collection retrieves details of all jobs and tasks every 15
seconds. Additional call stack and file usage data is collected
automatically for any job causing a "HotSpot", i.e. exceeding
user-defined limits for CPU percentage or Disk-I/Os. Data collection
consumes typically somewhat less than 0.1 % CPU.
2. Data Expansion and Analysis.
The collected performance data stored in compressed binary format is
expanded and analyzed. Any job suffering from or potentially causing
performance problems are kept in the
GiAPA data base for
further analysis.
3. GiAPA Reports.
A large number of exception and detailed reports can be requested,
often telling down to source statement number where to correct which
performance problems - use the REPORT menu
above to see
examples.
4. Additional Tools.
Should a performance problem require additional analysis, a number
of supplementary tools
are available to run more detailed job or program analyses.
|
