When executing a test, LoadView can abort test sessions and mark them with the Uncompleted status. In the most cases, this type of sessions does not present any statics that can be meaningful for the test result analysis. Thus, by default, we DO NOT consider uncompleted sessions as sessions with failures and DO NOT include uncompleted session data into the error statistics and test reports.

When the percentage of uncompleted sessions for a test surpasses 10%, statistics on these sessions are available. These statistics can be accessed by clicking the information icon next to the test status indicator in the Test Reports list and on the individual report page.

Generally, a rule of thumb is that if you have load test session details presented under the Log tab, there is no need for analyzing uncompleted session reports. Otherwise, under the Log tab, select to Include Uncompleted Sessions in the session log and check the session details for errors. If there are no errors detected on the test run, make sure if the test duration is long enough to replay the whole script.

There are two types of uncompleted sessions depending on their effect on the test results.

The first type is the most common and includes sessions that were started at the very end of a test and aborted due to the test duration time was over. Once the test duration time is up, LoadView waits 30 more seconds after the actual test end time and aborts all uncompleted test sessions automatically. If some test sessions were started at the end of the test run and not completed within these extra 30 seconds, they will be aborted. To sum up, sessions of this type do not have data that can affect the test statistics and can be omitted while interpreting the test results.

Example

Let’s say, a test session duration is 10 minutes and the test duration time is 30 minutes. If a session is started in the last five minutes of the test run, there won’t be enough time to complete the session and the system will abort it.

The second type includes sessions with script replay issues. In these cases, some or all test sessions will be aborted during the test run. While you may still have the test results presented in the Load Test Summary report, the test sessions log may contain statistics on fewer sessions than you expect or no statistics at all (if all test sessions were aborted). To troubleshoot tests with this type of uncompleted sessions, you need to resolve the issues occurred on script replay.

There are two main issues that lead to the abortion of the script execution:

  • Persistent errors that occurred in the beginning of the script execution and significantly increase the script duration. This type of sessions will be aborted on the script completion timeout (15 minutes).
  • Inconsistency between the script duration and test duration, when the test duration is shorter than the actual script duration. In this case, there will be no time enough to complete the script replay and all test sessions will be aborted upon completion of the test.

In this article, we will have a closer look at each type of uncompleted sessions and provide recommendations on how to avoid the related issues.

Sessions with Script Errors

Errors that occurred while replaying a script at the very beginning of the script execution are the second common reason for uncompleted sessions. For example, errors can occur due to slowdown under the load or script errors (e.g., not valid input data, validation failures). If LoadView failed to replay an action from the script, it waits for the action completion timeout (from 50 sec to 2 min depending on the action type) and moves forward to the next action in the script. Usually, the errors that occurred in the beginning of a script lead to the failure in subsequent actions. More failed actions you have in the script, more waiting time will be added to the script duration. Thus, when you have a long script, the script duration can increase significantly and exceed the script completion timeout (15 minutes). In this case, LoadView will automatically abort the session.

If the issue persists during the test run, all sessions will be aborted and the test results won’t be correct. To troubleshoot issues like these, review uncompleted sessions details and performance reports, and fix the corresponding script errors.

Example

Let’s say, you’ve created a script with a login step in the beginning and several navigation actions followed by keyword validation after each navigation. If the system fails to replay the login step in the beginning due to wrong login data provided in the script, all subsequent actions will generate errors.

Supposing your script duration is 2 minutes and you have 6 navigation actions and 6 related keyword validation actions recorded. The system will wait 120 seconds (navigation completion timeout) on each navigation to complete the action and then jump to the subsequent keyword validation. Since the navigation step failed due to a wrong login data, the system won’t be able to find the keyword and will wait for another 50 seconds (validation completion time) to complete the validation. In this case, total waiting time is:

Waiting time = 7 * (120 second + 50) = 1020 seconds or about 17 minutes

This way the actual execution time of the script will exceed the timeout of 15 minutes and the system will abort the sessions with the described errors and mark them as Uncompleted.

Long Sessions Uncompleted Due to Short Test Duration

One more reason for test sessions being aborted is a long test script, while a test duration is being too short to replay the whole script. To avoid this type of uncompleted sessions in your test, make sure that your test duration is enough to complete the script replay. To do this, configure the test execution plan to have the Test Duration value at least 5 times bigger than the actual script duration:

Test Duration = Script Duration * 5

Sometimes the actual duration of your script may be unclear since some steps such as keyword validation or delays between steps can increase it. Moreover, LoadView automatically adds user delays to simulate real user behavior. To find the actual duration of your script, check your test device validation results.

Uncompleted Sessions and Load Test Error Statistics

Uncompleted sessions do not affect test error statistics. However, you can include them in load test reports if it is necessary.