09 Jan 2020 12:02 AM
Hello,
I'm trying to monitor a program which would execute over 1hr and have more than 10,000 PurePath nodes. (In fact, the program be estimated as 4 hrs or more execution time.)
However, a single PurePath has the limit so that I can't full capture the PurePath.
I've tried to add entry point by setting sensor configuration as "active and start PurePaths", but according to documentation, it will "start a new PurePath if not yet started".
Therefore, I still can't full monitor this program.
May I have any way to force the purepath to start one or more new PurePath(s) by some rules?
Thanks for the help!
Solved! Go to Solution.
09 Jan 2020 08:13 AM
This is a common challenge. The only way I know to solve it is to start PurePaths at a lower point in the call stack. For example, not starting at the main method, but at a "processRequest" or "doNext" type of method.
Also, note we generally do not get much value from monitoring long running batch processes. Our best value is from monitoring user-initiated transactions where a real user is impacted if the transaction fails or takes too long.
21 Jan 2020 06:39 PM
Thanks for your help! It's helpful.
10 Feb 2020 07:19 PM
Hello,
I have a new problem which is relate the topic.
I need to split a PurePath to two smaller PurePaths.
However, the PurePath is start at "doFilter" which auto sensor set as default.
I still need the auto sensor settings so that I can't adjust it
Do I have any other way to deal with this problem?
Thanks for your reply!
11 Feb 2020 09:53 AM
This comes up from time to time. There is no way to "stop" one PurePath and start a new one. A sensor can either start a PP if none exists or append to one. Removing our header to purposely break the PP is an option, but unlikely it would be easy to accomplish.
11 Feb 2020 07:58 PM
Hi Dave,
Thanks for your reply!
I'll keep trying. Thanks.