In this video, we are going to review an end-to-end transaction. The first thing I am going to do is to go over and open up Precise. I like to do ‘Run as admin’. And, that opens up the Precise GUI [graphical user interface]. Here, I will authenticate to Precise. And, log in. The application in question is called ‘Banking’. And, it is comprised of an Apache and two WebLogic clusters that eventually talk to Oracle. And, so, a pretty complicated environment. And, I am going to click here on ‘Web transaction flow’. So, we are going to look at the end-to-end transaction. The default is for Precise to open up and show eight hours worth of activity. We can show near real time by clicking on twenty minutes. But, what I would like to is open up the last two days. So, I just come to this screen. And, specify two days And, that will paint up for me. Now, I am actually interested in the time frame over here. So, I am going to zoom in here by just selecting that time frame. One of the nice things about Precise is that you can zoom in and zoom out all dynamically. And, Precise will automatically adjust and show what had changed on the system. Here, I have the time frame of interest.
And, there is actually quite a bit going on the screen. Over on the left side, you can see that we have two transactions. And, you see the top number. And, that is the average response time. Precise gives you the ability to view the information different ways. So, here we are looking at average response time in seconds. So, that is a little bit less than a third of a second. I could do weighted. I could do average. I could ‘Executions’ to find out what is the most frequently accessed transaction in our system. It is not a big deal when you have two transactions. But, on a busy production system, you are bound to have plenty more. And, so, that is very nice. You can control the way that the data is displayed and presented. Precise always sorts in descending order to bubble up to the top resources to the top of the list. Now, here we have a profile, a stacked bar graph which shows us where the time is spent. And there is key on the right-hand side. And, you can see that dark brown time maps to database time. And, I actually had made a change in the system. And, you can pretty well see exactly when that change was implemented. And, what that change was very simply that Precise identified a hotspot being this particular SQL statement.
So, what makes Precise an APM [application performance management] tool is we show application code. Whether it is EJB [Enterprise JavaBeans], JVCs [Java View Controller] or servlets we show the underlying code, methods, and classes, etc. And, so here it is a particular SQL statement. Now, our DBA [database administrator] loves parallel execution. But, we saw that were filtering here by the ID column. Went ahead and did the test. Did an index on an ID. And you can see that the performance is markedly better from both the database time and the transaction time. One of the really nice things that Precise does is it puts things in context. And, what does that mean? Well, let us take a look at this top transaction. When I basically unfold, what it did as it traversed the transaction stack, I can actually see the application code that was executed. I can see that we were making an EJB [Enterprise JavaBeans] web service call out to a second WebLogic cluster where these SQL statements are in turn executed. And, by connecting the dots to the resource consumers, the drivers of activity, and at the database level that is the SQL statement, I can allow the DBA [database administrator] to see exactly what SQL statement that they need to focus in on. And, the reason that that is important is that if a DBA [database administrator] is just looking at things from a database or instance health perspective, they may not know about the particular SQL statement that is holding up a transaction. Now, this transaction was not slow. But, it had the capability of being a lot faster. So, what we are going to here is zoom into the after time frame so we can see the new execution time. And, it is less than two-tenths of a second. So, it basically dropped from a third to fifteen. So, it basically cut in half by one simple change in the system. You did not have to go out to buy a new server or that type of thing. I changed the application to run more efficiently. Now, that is a very simple example. But, I wanted to keep this simple. I want to also show you how on this transaction screen with the findings report we can often easily see in one screen exactly what needs to be changed in your system. It is actually that simple. So, thank you very much for your time and interest. I hope that you enjoyed the video.