OCR - transaction did not complete (give up)
Hi,
I created test script for Chrome browser but I get "Transaction did not complete - give up" when WaitForTextOCR function is used.
Below is a part of the script:
ResetRect rectIn
rectIn.Left = 110
rectIn.Top = 125
rectIn.Width = 90
rectIn.Height = 25
SleepFor 3, "Sleeping"
AddTransaction "03_alert_screen_open_ch"
ShowStatusBox "Loading Fraud Module Page", True
Window("Application=CHROME.EXE Caption='Page Title Undefined - Google Chrome'").Attach
Window("Application=CHROME.EXE Caption='Page Title Undefined - Google Chrome'").Type "{Return}"
StartTrace "03_alert_screen_open_ch"
WaitForTextOCR Window("Application=CHROME.EXE Caption='Alerts analysis and investigation - Google Chrome'"), "Alerts", True, False, rectIn, rectOut
StopTimerAndTrace "03_alert_screen_open_ch"
When I run the script by hand from the Agent Recorder everything is OK but when the script is running by the Agent Manager scheduler I get the mentioned error.
Do you have any idea what can cause it?
Thanks.
Jakub
Best Answer
Hi Jakub,
For troubleshooting you can add a call to ShowTextFromOCR before WaitForTextOCR to see what text the OCR engine is able to recognize during runtime. It will bring up a dialog with the list of words found. You may need to add a sleep before the call to make sure the window in question is fully loaded. You will also need to close the dialog manually but it will allow you to compare what is found during a scheduled run and a manual one.Also, if you have room to shrink the rectangle around the desired text that may help.
If you are unable to get it working you can open a support ticket, Go to page: http://support.dynatrace.com and click on the link Access the Support Ticket System to create a support case.
Best Answer
Hi Benjamin,
thank you for your reply. Today, we resolved the issue with Marty via support ticket. Not updated framework to 12.3.3 was the reason. After update everything works well.
Regards,
Jakub
Hi Jakub,
when you run the script from the Recorder, it will run in the logged-in user's account.
When the script runs in the scheduler, it runs in the Synthetic Agent's service's user account. Could that make a difference in your case?
I'd also try Marty's idea to see what the difference is...
