Improve Operation and Maintenance with IBM Z Common Data Provider
IBM Z Common Data Provider can help you realize intelligent operation and maintenance.
By Qian Xia Song, Ai Ping Feng and Ya Nan Tian06/24/2020
CICS® operation and maintenance in particular is of the utmost importance, and CICS performance data is vital in the daily use and maintenance of CICS. CICS performance data provides detailed CICS transaction-level information, such as the elapsed time of processors and transactions, and time spent waiting for I/O. For each transaction, CICS writes at least one performance monitoring record. CICS performance data also provides resource-level data which can be used for accounting, performance analysis and volumes planning. Therefore, enterprises are very concerned about whether CICS transactions are normal or not.
IBM Z Common Data Provider (ZCDP) provides SMF_110_1_KPI data stream to help you collect CICS key performance indicators (KPIs) to conduct problems analysis, performance evaluation as well as rational resources management. With ZCDP’s help, you can also customize data streams to collect. ZCDP can provide you with different data, and define data output and data operation rules to meet your business needs. All of this data can be made available through one single collection. Now, let’s look at how ZCDP can help you analyze CICS performance data and realize intelligent operation and maintenance.
ZCDP and CICS PerformanceIn their daily work, system administrators need to make a comprehensive analysis of current CICS performance after some specific PTFs are deployed. You need to know clearly whether the system runs normally, whether transactions response is normal, what indicators change and how these changes will impact the subsequent operation of your system.
CICS system problems are reflected on abnormal transactions data first. Generally speaking, before you deploy a PTF, you can collect SMF_110_1_KPI with ZCDP, and set up the performance benchmarks for PTF deployment. To make sure that the benchmarks are valid, you can collect data over different periods. After the PTF is deployed, you can use ZCDP once more to collect CICS performance data, and make a comparison between the data and performance benchmarks that you set up before the PTF deployment. The dashboards of analytics platforms, such as IBM Z Operations Analytics, Elastic Stack and Splunk, provide an intuitive interface through which you can have a clear view of performance differences before and after system maintenance.
You can also collect CICS KPIs with ZCDP to monitor CICS system resources use. From daily CICS performance data, you can get the model of how CICS applications use CPU and memory through machine learning, and in the meantime, monitor the overall CPU and memory usage of the system. If exceptions occur when system CPU and memory or a CICS application uses resources, ZCDP will give out an early warning. System administrators can check the dashboards of analytics platforms to find out the exceptions, and browse the system and application log to track and deal with problems. ZCDP can help system administrators realize intelligent system operation and maintenance.
SMF_110_1_KPI Data StreamSMF_110_1_KPI data stream provides 28 KPIs for CICS transactions, including:
- CICS transactions start time, stop time, elapsed time, name and statistics. If you need to set up algorithms to charge users for the use of resources, you can use this type of data collected to update the expense information in your organization’s accounting programs.
- The total number of syncpoint requests that are issued by the user task in a certain period of time (interval), the elapsed time during which the user task waited for input from the terminal operator after the user task issued a RECEIVE request, and the elapsed time of the transaction. The total processor time during which the user task was dispatched by the CICS dispatcher domain on each CICS TCB under which the task ran and the total elapsed time during which the user task was dispatched, etc. You can use this information to deploy CICS tasks and adjust CICS performance.
Define SMF_110_1_CUST Data StreamIn addition to the 28 CICS KPIs collected by SMF_110_1_KPI data stream, CICS performance data provides detailed CICS transaction-level and resource-level information. You can define your own data streams according to your needs, and collect more performance data.
Please note that, in the IBM Z Common Data Provider User Guide for ZCDP 2.1 version, there’s a topic called “Streaming key performance indicators for CICS Transaction Server for z/OS monitoring.” This topic describes in detail how to collect KPIs to monitor CICS TS, including how to create DEFINE TEMPLATE statement to customize the data streams, how to define SMF_110_1_CUST, and how to update your development platform to ensure that the data is properly streamed to the open end.
CICS performance monitoring records data is of vital importance to the maintenance and monitoring of a CICS system. With the help of ZCDP to collect CICS performance data, you can collect customized data according to your business needs, conduct different calculations and make the operation and maintenance of your CICS system intelligent. ZCDP monitoring features include:
- Monitoring transactions response time: For example, you can use INTERVAL (START, STOP) to calculate the elapsed time it takes for a transaction request from starting to being processed.
- Monitoring processor usage: USRCPUT_TOD records the time that transactions use CPU. You can calculate the average CPU usage with SUM(USRCPUT_TOD/4096E6)/RECORDS.
- Monitoring storage usage: For example, SUM (SCUSRHWM) provides the maximum amount (high-water mark) of user storage (bytes) allocated to the user task below the 16 MB line in the user dynamic storage area (UDSA). SUM (SCUSRSTG) provides storage occupancy of the user task below the 16 MB line in the UDSA in kilobyte seconds.
- Monitoring volumes and throughput
- Monitoring CICS resource use
- Monitoring files and databases usage
Get more information about ZCDP.
Qian Xia is a software engineer in IBM Systems. She has been working in IBM Z for eight years, and has rich experiences in the product features development and customer problem solving of IBM z/VM and IBM z/OS. In the last three years, she has been working on IBM Z Common Data Provider as a technical lead.
Ai Ping is a software engineer in IBM Systems. She has been working in IBM Z for nine years, and has abundant experiences in the product features development and customer problem solving of IBM z/VM and IBM z/OS. Since 2017, she has been focusing on the development of IBM Z Common Data Provider.
Ya Nan is the information developer for IBM Z Common Data Provider, with rich experiences in designing and developing technical content.