Hi All,
We have been having an issue recently where a job that previously had been running well and efficiently is now starting to take exponentially longer to complete activities the longer the job runs. We have been able to watch the job via a stream start off with where activities process at the correct pace but after about 20 minutes of running, certain activities dramatically slow down. The average transactional runtime increases from about a minute to sometimes upwards of 18+ minutes. There were no changes made to this automation within the last 3 months but this randomly started happening about 2 weeks ago now. We found that when the browser is closed and reopened the transaction time drops back to normal for the next few transactions then continues to steadily rise again.
We implemented a trigger that restarts the job every 30 minutes so that it manufactures that browser restart before the transactions become unmanageable. Below, I’ve uploaded a screenshot:
- One displaying the activities themselves and their max and min elapsed time.
- The other, a supporting image to the activity timing.
- The last showing the average transaction time by date and how this is a massive jump overnight on May 9th.
Another thing to note is in the above image, all activities have time limits set:
-
The longest activity is a simple click action that if it cannot find the element on screen it should fail after 30 seconds. The timing is all set to the default with no delays.
-
“1 of 1” has a 15 second limit where it should no longer attempt to look for that on screen but as you can see, it is attempting to find that at times for upwards of 2+ minutes or longer if the automation keeps running.
I will try to post the other screenshots described above but I cannot post more than one embedded image since I am a new member.
Thank you for any help or suggestions!
