I don’t know if you solved this, but I just spoke to a guy from IA and he mentioned that you can see what the pipeline is doing from the thread stack trace. So if you get a thread dump and look at it in Kindling, you might be able to work out where it’s stuck. Filter the dump by “alarm-notification” and you should see results in there.
But a throttled alarm pipeline is generally caused by something delaying it from exiting, like:
- alarm consolidation
- long-running scripts in the pipeline
- loops in the pipeline that haven’t exited yet