I ran a workflow that opens an Excel file with 300 rows in a specified folder, reads search words starting from the 5th row downwards, performs web scraping against SaaS (Software As a Service), and outputs the results back to the original Excel file. However, I noticed that at row 60, the results from the web scraping could no longer be obtained in a variable using the NGetText activity, and I suspect that the web server became overloaded, leading to the inability to navigate the screens. Someone suggested that closing and reopening the browser tab or restarting the browser every certain number of rows might improve the situation. Currently, I have only one browser instance running; could there still be a chance for improvement under these circumstances? I would appreciate any insights from those who are knowledgeable about this.
Quick question - does the Web UI require you to login? or is it a public webpage?
If it’s public, I would suggest opening and closing the browser after each record, and when opening, try using an incognito (Chrome) or private window (Edge) tab each time. This will prevent Cache buildup and session instability.
If there is a login AND navigation involved, I would the suggest using something like a “Check App State” activity before you read the text from the webpage each time, and perhaps add a small delay before (1/2 seconds) reading the text to make 100% sure that the text is in fact there and that the browser is loading correctly after 60 records. One last thing is adding a retry scope around the “Check App State” which after 3 attempts will refresh the browser, or close and open again from the beginning. You need to add as much stability as you can for handling these types of scenarios.
Hi @JarrydScott
Thank you for your reply.
I do not need to login by workflow because I manually loginned this system.
I have already used Ncheck state activities as well as delay options.
So, the most part of your advice is useless
Hopefully I expect your direct answer to my question
Firstly, I would recommend upgrading to a later version of Studio. version 2022 is quite old.
Secondly, give my first option a try of “closing and opening the browser per record.”
I also see you have created a new post with an extended version of this question. You said at the end “In this case, it is unlikely that the web server side will be overloaded, so there is a high possibility that it will work if the PC is restarted and run due to the local PC side.” So try the first two options and then restart the PC and give it a try.
