Execution logs are the logs which got captured during your test based on client server communication.
During on of my project .SWT file was taking longer time.
Response from Team:
Very 1st point generally we should not keep login part in Action section of script as it doesn't seem to be a realistic scenario. Think does real world user (Customer) login and logout that frequently as you are making your application to do in your scenario? My suggestion would be to keep login in init and logout in end section of scripts. But still taking your scenario into consideration.
I am not sure if this way I can help you without having look at the results but still you can try as mentioned below:
1. You must be having .lrr file as test results. (If test was executed in controller)
2. Open .lrr file with Analysis LR component. Let it open complete 100% data. Data generation can be seen at bottom on analysis component.
3. Once 100% data is generated. Open Average Response Time graph.
4. Right click on graph select merge graph and select running vuser graph. This will give you running vuser and average response time graphs in single graph.
5. With this graph you can find exactly after what user load login transaction started taking more time. Once you find the point where login transaction starts consuming more time note down the time and note down what is user load at that particular point of time.
6. Now open other graphs like CPU utilization, Memory utilization (Committed Bytes/ Available MB) check there consumption level if there are normal utilization if you see utilization beyond 90% note down user load and point of time when your server starts consuming beyond 90% level.
Report this to development/Application team with this detail. This will help them.
Ask them to monitor application server from back end while you are running load test. This will give them great understand of root cause.
7. If you have tools like NewRelic,Appdynamics or Dynatrace from there you can drill down transaction wise and find out the culprit behind the high respone time. It could be a database query taking high response time,Network delay, high Disk read/write or multiple queries getting fired for the same transaction.
No comments:
Post a Comment