Performance Guidance

To achieve and maintain the best possible performance for your IBM® z/OS® Connect Servers, you can tune the environment and z/OS Connect settings to match the requirements of your application workloads.

For z/OS Connect to perform well and to meet agreed service level objectives, the performance relies on several key resources such as CPU, memory, network, and storage. If you believe that a performance issue exists with z/OS Connect, then ask the following questions to help identify the cause:

Why do you suspect a performance problem?: Are response times slower than expected?; Are timeouts occurring?; Are more resources being consumed?
Why do you suspect the problem is with z/OS Connect?: Are requests taking longer than usual in the server?; Is garbage collection taking too long?; Started task Was the z/OS Connect address space that is swapped out?; Has the JVM run out of memory?; Started task Do you see error messages in the JOBLOG?; Do you see error messages in the messages.log?
What changes occurred before the performance issue?: Has an update of z/OS Connect been applied?; Has a configuration parameter been changed?; Have the characteristics of the workload changed?; Have other products in the workflow or underlying stack been updated?
How often has the performance problem occurred?: On multiple occasions?; Regularly at a particular time of the day or day of the week?; Just once?
What is the scope of the problem?: Limited to one API, service, or API requester?; Limited to one z/OS Connect Server?; Started task All z/OS Connect Servers on a single LPAR?; Started task All z/OS Connect Servers on all LPARs?; Other products as well as z/OS Connect and are these products upstream or downstream from z/OS Connect or both?

Check the What's new in IBM z/OS Connect? for problems that are already fixed in z/OS Connect and for the minimum version of z/OS Connect. For known performance issues, see Performance considerations.

If the problem is still unknown, z/OS provides a wealth of tools including SMF, RMF, and monitors to help diagnose performance issues. These tools are discussed in the following topics.