I'm working on with and EDA v126.96.36.1992 via Atlas and modifying a syslog parsing trigger, but suspect this may be an issue with any trigger. As I make modifications / fixes / additions to this trigger and save the changes, I notice that versions of the trigger prior to the save are continuing to execute on the EDA as if stuck in memory or tied to a flow. Though I see this behavior carry on for some significant time (in excess of 30 minutes) post the saving of the trigger.
As an example, there was an exception being thrown on Line 177 of this trigger. I added a line to attempt to reduce the exception, which moved "excepting" line to 178. Once I saved the trigger, the runtime log reported exceptions on both line 177 and 178 for the same error.
I went back, added empty lines to 176-179, saved the trigger, and the resulting runtime log output was three instances of the error reflecting to lines 177, 178 (from the first change), and 181 (newest line # for the erroring code).
It wasn't until I stopped and restarted the capture that the legacy code discontinued execution. While this is a short term fix, long term, having to restart excap after every trigger modification is not good practice.