Description
Symptoms
The incident record is closed, but this state transition is not showing in the activity log or audit history. As a result, relevant SLAs are not transitioning stage and are still showing as "Paused" (when they should be "Completed")
Release
Jakarta Patch 7
Cause
Two after Business Rules with the name execution Order are why this intermittent behavior is happening (depending on which of the two runs first).
Resolution
The user has the following after Business Rules running:
- incident autoclose
- Create Problem when state is Resolved
Both of these after Business Rules have an execution Order of 100. This is the issue, and why this behavior is intermittent.
Looking at INC0010000, the incident was moved to a state of "Resolved" at 2018-07-13 22:11:47 by James.
Similarly, per the user's Business Rule "Create Problem when state is Resolved", PRB0010000 was opened at 2018-07-13 22:11:47 (the same time).
Unfortunately, within Business Rule "Create Problem when state is Resolved", the user has the following snippet of code: current.setWorkflow(false); This is not bad in-and-of itself. It is only causing an issue because the two Business Rules are racing to see which one will execute first (again, hence the inconsistency of the behavior).
The above piece of code will prohibit all Business Rules from executing after it runs (in this case, it prevented "incident autoclose" from running), which is why the moving of State to "Closed" was never tracked, and thus, never passed down to the task_sla records to progress them to a "Completed" stage.
Again, the issue here is that both of these Business Rules, "Create Problem when state is Resolved" and "incident autoclose" are executing on the same Order # (100).
Suggest that the user modify their "Create Problem when state is Resolved" to execute on a higher order number than "incident autoclose".
Making the above modification will ensure that the "incident autoclose" Business Rule runs first and succesfully passes down that closure to the task_sla records, resolving this intermittent SLA behavior. Then, the PRBs will be safely created after "incident autoclose" has already run.