Suddenly, job is randomly/sporadically failing -- but works 100% of the time if I RDC into job before it fails

Help Robot
, , ,
1

I’m a bit at a loss to what is going on. Suddenly, some of my long-standing jobs (have been running daily without much issue for a year or more) are faulting on the initial activity involving web browser (chrome) activities. It will be on the first activity involving looking for an element (e.g. Input for username).

I first thought some Windows popup/process was overlapping the browser and causing the interruption.

I have tried quite a few things. Adding at the start activities such as Focus, Activate, kill process for chrome.exe to end any background Chrome stuff that might somehow interfering, throwing them into a Retry Scope, different order of the above activities, etc.

Sometimes the jobs work fine and I think I fixed it, sometimes it fails in the same exact way. It doesn’t matter which activity I have first involving looking for elements in the browser; if it is going to fail, it will fail on whichever is first.

While I have these activities in the Retry Scope, with logs I can see that it is going through this loop to keep trying. If I access via RDC at any point, the job will immediately continue without any issue and complete the job. When I remote in, there are no popups or anything I can visually see to indicate why the browser is not able to be in focus prior to my arrival.

The errors I get depend on which activity is first for it to fail on, but such errors as:

“Message: Cannot bring the target application in the foreground.”
“Message: Could not find the user-interface (UI) element for this action.”

And again, this always works without issue locally and works 50% of the time as unattended via orchestrator. And it used to work 100% of the time. Only in the past week have I began to encounter this issue on multiple jobs.

2

Hi @mjconns

This is a very known issue , please try to use Check App state activity(part of Modern Activities) , and add more amount of delay value in it and check .
Also you can try enabling Remote Debugging in Studio, and check what is going on with the process in UnAttended mode .

Hope this helps you in right direction and your process works smoothly .
Please refer this links below for reference

3

@Kanak_Lata_Singh, thanks for the ideas. I couldn’t find posts or documentation to my exact type of issue, so I did not know it was very known.

As fate would have it, today, nothing is breaking. Things are back to normal, like they were a week ago, and everything working as normal. So, I will see about adding in the suggested activities, but I’m not sure if I will see any difference with your ideas or not since the bot is behaving now.

4

Hi @mjconns , are you using a CV activity? What sort of license do you use?

CV throughput (megapixels/minute) is limited based on the license type.