Overview of Non-timing CICS Transaction Metrics

Overview of several types of non-timing transaction metrics captured in CICS Transaction data.

More CICS Transactions Videos

1. CICS Timing Metrics – Overview
2. Transaction Response Time Analysis
3. Timing Analysis by High-Level Components
4. Response Time Analysis Scenario 1 Values Differing by System ID
5. Integrating WLM Performance Index and CICS Transaction Analysis
6. CICS Web Services Metrics
7. Overview of Non-timing Transaction Metrics

Video Transcript

All right. So now let’s move on to the data request subgroup or metrics pertaining to CICS interaction with other resource managers are captured. So here is the rate of Db2 SQL and IFI requests by transaction. And again, we could view this by web service name we chose to. So other metrics in this group is the volume of IMS requests, the volume of MQ requests, and then also here, we’ve got wait times for the resource managers. So, here’s the IMS wait time and then Waits for Db2 thread and Db2 connections to become available if there are any, is captured in this data.

All right, metrics on virtual storage usage at the transaction level for each of the many types of CICS storage areas are captured in this group. So, using extended user DSA, as an example, here we see the number of GETMAINs in that area, storage occupancy, and the maximum storage allocated in that particular storage area. Again, all at the transaction level.

Another category of data available in the CICS Statistics data is file activity. So for transactions that are doing their own file I/O and not using Db2, we see that activity by request type. So here’s the number of GETs and PUTs, BROWSEs, and ADDs at the transaction level. CICS is a major exploiter of the zOS system logger functionality, and the transaction data capture the volume of log stream rights. So if we look at this transaction with most of those, and interesting to also bring on the wait for journal I/O timing metric, and when we do that, bring them together on the same chart, we find a very close correlation between the number of log stream rights and the amount of weight for journal I/O, as we might expect.

All right, the program load group captures the frequency of various mechanisms for loading programs at the transaction level. So here’s a LINK and INVOKEs, here’s LINKS of User Replaceable Modules, XCTLs, and program LOADs. So we see a high volume of program load activity for this top transaction. And we go over and look at the load time timing metric. We can see that for that transaction it does translate into a measurable but small amount of wait time per tran. The other category is temp storage, different types of activity here. Here’s the PUTs to Auxiliary Temp Storage by transaction, here’s to Main Temporary Storage, and then GETs and PUTs, GETs To and PUTs From Shared Temporary Storage.