Average SNA response time measured over the last response time collection interval. Response time collection interval is specified on the AvgSampPeriod parameter of MONITORGROUP statement in the TCP/IP profile dataset. The SNA response time is calculated by subtracting the Average IP Response Time from the Average Total Response time. The most likely causes of
a high average SNA response time are an application workload spike or z/OS
system resource constraints.
- If the cause is a spike in application workload (for example, the number
of currently active connections is a high percentage of the high water mark
for the number of currently active connections), consider creating an additional
instance of the application.
- If the cause is not a spike in the application workload (for example,
the number of currently active connections is not a high percentage of the
high water mark for the number of currently active connections), then determine
whether system resource constraints exist.
To further investigate system resource constraints, use the z/OS
Resource Measurement Facility to evaluate workload delay metrics such as CPU
and I/O for both the application and z/OS Communications Server address space
to determine if possible system resource constraints exist. This situation
is triggered when the average total response time is greater than .5 seconds.
As provided, this situation is not started automatically.
Note: Applicable only when monitoring TN3270 servers running on z/OS v1.8 and higher.
|