Performance information available

To monitor the performance of your IBM® z/OS® Connect servers and to diagnose performance issues, z/OS provides a wealth of tools such as monitors, SMF, and RMF.

You can monitor the performance of an active IBM z/OS Connect server by using various products such as IBM Z® Omegamon for JVM, and the Java™ Health Center. RMF can be used for both real-time and historical data, while SMF records are used for recording system usage over a period.

Omegamon for JVM

Omegamon for JVM provides generic JVM monitoring capabilities plus real-time data about IBM z/OS Connect APIs, services, and API requesters. It provides detailed data on individual requests, including status codes, timings and payload sizes, plus summary views. You can use Omegamon for JVM to set alerts for various reasons such as raising awareness and taking corrective action for potential performance issues. For API provider requests it provides data on in-flight requests.
Note:
  1. API provider workloads are supported in Omegamon for JVM V5.4 and later.
  2. API requester workloads are supported in Omegamon for JVM V5.5 Fix Pack 2 and later.
For more information, see IBM Z OMEGAMON for JVM.

Java Health Center

You can also monitor the JVM that your active IBM z/OS Connect is running in through the Java Health Center (part of IBM Monitoring and Diagnostic Tools). This monitor provides real-time data for the JVM including Java heap usage, data on garbage collections, and thread usage, with minimal processor usage (less than 3% CPU).

The Java Health Center is discussed further in the topic Threads, Java heap, MEMLIMIT, REGION.

For more information, see Health Center in the IBM Monitoring and Diagnostic Tools documentation.

SMF

SMF (System Management Facilities) is a component of z/OS that collects and records system and job-related information. SMF records are useful for recording system usage over a period with minimal overhead.

For IBM z/OS Connect , SMF 72 and SMF 123 version 2 records are especially useful. You can use the SMF 72 records to review the CPU usage of the IBM z/OS Connect address space. If the workload is classified in z/OS WLM, you can also use SMF 72 records to view the CPU time that is used for one or more APIs, services, or API requesters. As IBM z/OS Connect is primarily written in Java, up to 99% of the CPU is potentially off-loadable to zIIP processors. The GCP and zIIP usage for the server can be monitored by using SMF 72 records.

SMF 123 version 2 records provide details of each API, service, API requester, and administration request, and include HTTP status codes and timestamps as requests enter and leave the IBM z/OS Connect server. These provide invaluable information that you can use to see how much elapsed time is spent processing requests inside the server, the time spent waiting for responses from the SOR or API endpoint, or how much time is spent obtaining an access token from an authorization server with API requester.

There are several other SMF records that provide useful information about your IBM z/OS Connect server and might be useful to diagnose a problem where IBM z/OS Connect is part of the workflow. For example,
  • SMF type 30 CPU and other information for each address space
  • SMF type 70 CPU usage information for z/OS
  • SMF type 72 z/OS WLM (workload management) information
  • SMF type 102 Db2® information
  • SMF type 110 CICS® information
  • SMF type 115 IBM MQ information
  • SMF type 119 TCP/IP information
  • SMF type 120 Subtype 11 Liberty information
    • Provides CPU data for a request (WLM classification must be configured)
    • Enable by configuring zosRequestLogging feature in server.xml
Note: IMS does not write SMF records. The equivalent data for IMS is written to the X'01' type log records.

RMF

RMF (Resource Measurement Facility) is an optional z/OS feature that can be used to record and monitor z/OS performance, both historically and in real-time. RMF includes an ISPF application to view this information in real-time. RMF also writes SMF 70-79 records for the generation of reports for historical data containing valuable performance information including CPU usage of your IBM z/OS Connect servers and optionally the CPU usage of one or more APIs, services, or API requesters. For more information, see RMF monitors and post-processor.