CICS Transaction Class Metrics

This video explores several metrics that provide visibility to help sites effectively utilize the CICS TCLASS construct to manage the mix of transactions in a region.

More CICS Statistics Videos

1. Assessing CICS Storage Health
2. Assessing CICS Transaction & Enqueue Health
3. CICS Transaction Class Metrics
4. CICS TCB Dispatcher Metrics – Part 1
5. CICS TCB Dispatcher Metrics – Part 2 (T8 TCBs)

Video Transcript

So now let’s explore a few other CICS Statistics metrics. In Transaction Class metrics, the number of times reaching a T class limit, right. Transaction classes are a means to manage the mix of transactions, to protect against one type monopolizing a region. And here we can see for this particular CICS group that we’re reaching the limit frequently. So, let’s look and see which transaction class that’s happening in, and it’s in these two transaction classes, the one that ends in WREV and the one that ends in GYEV.

So again, we want to look at that now that most frequent one across the regions in the group. We can see that it’s occurring across all the regions in that particular group. We might also, again, back to the question, is this the common behavior or an anomaly, we can compare it with the previous day. And as we do that we see that it is a common behavior.

All right. And then finally here, let’s go ahead and see what the max value for the T class is. So let’s go ahead and add that on this chart as well, of course, on a different axis because it’s a different unit to measure. When we do that, we find out that the T class value is set to five for this particular transaction class.

Okay. So having seen that we’re hitting that T class constraint several thousand times an hour. So that might lead us to ask what’s the percentage of transactions that are being delayed. So again, let’s look at that group and look at it by transaction class. So for that particular T class about 18% are being delayed. And for the other one that happened less frequently, actually, a larger percentage, about a third of those transactions are being delayed for that particular transaction class.

We want to look at this across time, we can see that the percentage of delayed transactions, usually something during prime time, is around 20%. And then during that one peak time, it was 33%. So that might lead us to another question. And that is for those queue transactions, what’s the average time spent queued. And fortunately, the CICS statistics data gives us that as well.

So again, let’s go ahead and look at this data by transaction class. So again, that transaction class was hitting the limit more frequently, but it was a lower percentage. The average time queue is also lower. It’s usually less than a second. Although at that peak hour, it was five seconds. And again, those values are lower than they are for that other transaction class. So if you’re using the T class mechanism to manage the mix of transactions in your regions, then the statistics data can give you great visibility into how that’s operating.