|
GiAPA
standard
performance data collection retrieves the following data (a total of
28 fields) every 15 seconds for all active jobs and tasks:
•
Job or task
name, user name, job number
•
Job type,
priority, pool, date and time
Total values (from
start of job or task) for:
•
CPU
milliseconds
•
I/Os
(physical/logical, synchronous/asynchronous, read/write, data
base/non-DB, miscellaneous/perm.write)
•
Communication puts and gets
•
Exceptions
(decimal, binary, floating point, extended address overflow)
•
Transactions
(number and response time)
•
Print lines
Additional
information collected for “HotSpots” when job exceeds user defined
limits within a 15 seconds’ interval:
•
Call stack,
including statement number and library of last called application
program
•
Detailed
statistics for all files opened by the job.
Data collection
I/Os are minimized by using binary format variable length records
compressed into spanned blocks. This is why data collection
typically uses less than 0.1 % of one CPU. |
