20-12-25 05:25 PM
📌 Title:
Warning when attaching to browser causes process to hang
🔍 Description of the Issue
We have a couple of processes using the same VBO that intermittently get stuck in a warning state in Control Room when attempting to attach to a Chrome browser instance. When this occurs, the process does not throw an exception — instead, it hangs indefinitely until a human intervenes by manually closing the browser on the resource PC. Once the browser is closed, the process resumes and continues successfully.
🎯 Expected Behaviour
The process should fail to attach, trigger the retry logic, and then continue execution without requiring manual intervention.
💥 Actual Behaviour
Instead of failing and following the retry loop, the process immediately enters a warning state on the first attempt to attach and does not progress. No exception is thrown.
Errors logged when the issue occurs:
Important:
This process is designed to fail to attach when the tab is unavailable, and normally does so correctly both before and after these occurrences. The abnormal behaviour only happens during this warning hang condition.
🧪 Steps to Reproduce
This issue is intermittent and has no identifiable pattern.
It has occurred 16 times since 09/12/2025, but cannot be reproduced on demand.
⚙️ Technical Details
🧩 Troubleshooting Attempted
❓ Additional Context
Any insight or suggestions would be gratefully accepted and considered.
20-01-26 09:06 AM
Have you managed to find a solution?
20-01-26 11:03 AM
No not yet we have tried several things to try to resolve the issue but with no luck. Our process controllers are currently having to manually intervene and close the browser when the process enters the warning state.
20-01-26 08:51 PM
here in our organization , what we are doing is , before launching the browser we are killing Chrome, iexplore, msedge and Outlook , and giving 10 seconds of wait time and then launching the browser.
you can try this out
21-01-26 09:03 AM
@naveed_raza thank you for the suggestion. I have seen this kind of logic in some of our processes but I will be sure to investigate the logic flow for these specific warning issues and ensure we are taking that approach in this instance. Again thanks for the suggestion.
21-01-26 09:28 AM
@naveed_raza I have had a chance to go through our logic for this process and I can confirm that when we initialise we first check the environment and kill and running processes including edge and chrome. We do have a pause of 2 seconds after killing chrome but I will stretch that out to be longer to see if it has any effect on the issue.
2 weeks ago
Hi Robert, did you ever find a solution to this issue?
And can I ask if it happened upon upgrading to 7.4.1?
A client of mine has a similar issue, where BP hangs during the Attach stage on 7.4.1, and we're having to manually restart BP (sometime closing the browser works too). We're looking into an automated restart - by a separate program - if the resource is in warning state
a week ago
Hi,
I never did get to a route cause of the issue but we took several steps that seamed to remedy the issue somewhat but unfortunately I couldn't say with confidence what the solution is. Yes it was after an upgrade to 7.4.1 but not immediately.
The things I tried were:
1. Clear out (archive) a load of session logs.
2. Restart the Blue Prism application server.
I am sure we did other things but I don't remember exactly what and have a memory that the above seemed to have a positive impact.
a week ago
Hi Robert
Thank you for contacting Blue Prism community.
Here is some thought maybe you find it useful:
The key thing to understand is that Control Room's "Warning" status isn't a distinct failure state — it's just a timer: if a session's current stage hasn't updated within the warning threshold, it shows as Warning. So what's actually happening is that the Attach stage's call into Chrome is genuinely hanging (not failing) — nothing is thrown, so retry/recover logic never gets a chance to run. Closing the browser forcibly breaks that stuck connection, which is why it immediately "resumes" and then throws the expected "Not Connected" on Detach right after.
Because this is a hang rather than an exception, it can't be fixed from inside the process — there's no error to catch. The onset date and Chrome version make an auto-update-related cause likely (Chrome's move to Manifest V3 suspends its extension service workers when idle, and a wake-up race at the wrong moment could cause exactly this kind of silent stall). Worth checking:
Since there's nothing to catch, the practical fix is an external watchdog: something polling BPASession for sessions stuck on the Attach stage past a threshold, which force-closes Chrome automatically — essentially automating what's currently done by hand.
Regards
Hossein Azimi (APJ team)