What are five best rules of thumb when reviewing output from system monitors for preventing future problems.
The system monitors could be any SMF/RMF MICS TSO/MON MXG OMEGAMON system monitor type report for reporting on the mainframe.
Example might be
1. sustained CPU utilization over 85%
2. Long I/O ques
3. large variation in on-line response time from off to prime shift . . .ect . . .
4. any utilization reaching 100% just before response times start growing
5. an under paid over worked system administrator with a tooth ach?
what do you think?