Highlighted
Absent Member.. Absent Member..
Absent Member..
114 views

Expert day question: Effective error reporting for large number of users and long runs

Why Loadrunner or performance center is not designed for error reporting based on complete iteration / transaction only for failed transactions ?

 

This is essential requirements to track  meaningful failures and take necessary actions to fix them. Enabling whole logging is meaningless as it slows down client process when users cares about all information about failures only. Just giving stacktrace of current call failures won't report what data is used in current iteration/ transaction if request won't use those parameter but its essential to find bad data when large number of users or test run for long hours.

 

In performance run settings log only on error doesn't give complete information on what parameters passed to iteration/ transaction?  What was response for previous requests or actions before the failures from beginning of the iteration ? This becomes difficult to track and related to server logs for finding root causes.

 

This is very simple to design to log all info when iteration begins and flush/commit to files only on failed transactions and errors if all transactions are success all such logs should be ignored at the end of iteration. This way user gets complete trace of errors and transaction failures for failures and don't have worry about successful transaction when there are ton of users and millions of transactions.

 

 

Thanks!

Sanjay

 

0 Likes
2 Replies
Highlighted
Absent Member.. Absent Member..
Absent Member..

Re: Expert day question: Effective error reporting for large number of users and long runs

Hello Sanjay
The detailed information of each vuser run can be found on the raw results, there will be a log file for each vusers, this of course will be according to the log level you set, you can set full with only parameters replaced of standard, the same way while running you can see the runtime data, taht will show you the used paramters as well. Notice that there standards for types of tests, usually a test that have the logs enabled is a debug test and not perfromance one, since the impact the loaggin process creates on perfromance is not desired, this is up to each vuser, but you can definitelly find the information you need on logs.

hope this helps
0 Likes
Highlighted
Absent Member.. Absent Member..
Absent Member..

Re: Expert day question: Effective error reporting for large number of users and long runs

I think you didn't get the point. Think about you are running test with million users and 3 days. During tests there are hundreds of transactions failing at different points and logs don't show what data/ previous request responses were only for failed iterations. As enabling extended/ standard logs will be over kill for all the users. We need extended logs only for failures. Let me know if you have experience with such tests and what options product currently supports. 

 

Looks like you need expand you vision beyond vugen and think what happens in real life during large number of users and long runs where current designed system doesn't throw useful information for log on error runtime settings. It just throws stack of failed call instead of complete trace of full iteration which is very important to mine server logs to find why transactions failed.

 

Thanks!

0 Likes
The opinions expressed above are the personal opinions of the authors, not of Micro Focus. By using this site, you accept the Terms of Use and Rules of Participation. Certain versions of content ("Material") accessible here may contain branding from Hewlett-Packard Company (now HP Inc.) and Hewlett Packard Enterprise Company. As of September 1, 2017, the Material is now offered by Micro Focus, a separately owned and operated company. Any reference to the HP and Hewlett Packard Enterprise/HPE marks is historical in nature, and the HP and Hewlett Packard Enterprise/HPE marks are the property of their respective owners.