Purepath Analysis and Filters
Hello All!
I have been looking on the contribuitors tab for a filter option that allow me to sort by API, classes or other metrics from that table.
However, I only see basic options of time that applies directly to the dashlet.
I am interested in this options since my purepath is too big to analyze on Dynatrace (50,000 lines in the best case).
Maybe I just don't know exactly where to look at 
.
Is there a way to manipulate this data? Or do I have to export it and parse it with external tools (ex: Excel, DBs?). Btw, I am using DT 6.0
Thank you for your comments!
If you are looking to view top contributors to a very large purepath, a helpful view is to drill from the very PurePath to the Method Hotspots view.
This will sum up all the same methods in a PP so you can see the total. For instance, if method doFoo() were executed 1000 times in the purepath, each at 20ms, it might be hard to spot that as a problem, especially in a very large PP. If you drill from the PP to the Method Hotspots, you'll see doFoo() executed for a total of 20,000ms very easily.
Additionally, in the bottom panel, where the PurePath tree is displayed, you can switch the tab from the PurePath tree to the API view. This will provide API pie charts for Total Execution as well as CPU execution. If you'd like more from the API view, drill down from the large PurePath to the API Breakdown view (right click drill down option). In Addition to Total Execution and CPU, this will show Sync and Wait execution, and Suspension execution can be revealed by right clicking on the headers to add it.
Another option is to drill from the very large PP to the Response Time Hotspots view. This view will show, broken down by CPU, Sync, Wait, Suspension and I/O, where time is being spent by both tier and API for a higher level. If you were to click on any section of a bar in any column, this will then bring you to the Method Hotspots view mentioned above for that section. If it's JDBC I/O, it'll take you to Database Hotspots.
Even after taking advantage of the "ANDing" of the find strings?
Yeah, I did that. the problem that I face is that I am analyzing batch jobs that runs procedures several times, since are reading records and doing the process again and again. Besides, those AND are good when you know what to look at exactly. In my case, I want to discover it.
Right now, I just move to export the data and process it on mysql.
Thanks anyway
Hi David
Yes, I have used those options. However, are still too generic. As I was saying, the sort will still give me tons of lines; same with the sort. Too many lines to actually be able to read correctly the data.
Have you tried using the "Find" filter on the contributors tab? Either use Ctrl-F or just start typing and you can enter a class name, package name or other string to filter down the contributors tab. You can also sort on any column by clicking the column header.
HTH,
dave
