Interpreting exec time at service method
Hi!
Please help me undersand the following case. I found that the exec time of the service() method was 1064.75 ms, however, as you can see on the below source code this invocation should have been around 0 ms because there were not any relevant code here. The service() method invocation was caught by instrumentation based servlet sensor (not the autosensors) so it should be correct timing information but it seems it is not.
Hi
I agree with you - this looks a bit odd. However - I am pretty sure there is someting else consuming the time that we are not capturing in the moment through Auto Sensors. Auto Sensors do a great job overall - but - for individual PurePaths it is possible that we do not capture all the details as this depends on the current load on your system and your auto sensor resolution setting. Can you export a purepath like this and send it over?
Andi
Thanks for your help again! Unfortunatelly I could not upload the session here, so If you give me your email address, I'll send it to you.
Check out my Share Your PurePath. It tells you how you can get in touch with me: http://bit.ly/sharepurepath
Andi
I wanted to update the community. Balasz shared his PurePath with me and I sent him an analysis. It was very interesting to see what is actually going on in their application. I think there are several lessons learned here that I summarized in this screenshot of a PowerPoint I sent him back. I hope many of you find this useful:
Hi Balazs,
Sometimes, expanding the Show all nodes option, upper right corner of PurePath tree, helps a little, just a humble idea.
Just saw Andreas reply now, after posting, really interesting analysis indeed!
