Suspended definition starts with signal

Hello,

I have suspended a process definition that has a signal start but sometimes it gets started.

I couldn’t find in which cases it starts, it seems to happen randomly and I cannot reproduce it as the most of times it doesn’t start.

In my code I have a rabbit listener that listens to some event and calls:

runtimeService.signalEventReceived(eventName, variables);

The definition is suspended for a long time so I guess that the issue is not that there is some message stuck in the queue of rabbit or at some flowable db table and waits for execution. Also, I don’t know if it’s a cache issue there are cases where I trigger the event for the signal multiple times and only one time it seems to run the instance.

I have found an old similar issue: Process instances can be created from suspended process definitions via signal but as I understood and saw from the code running, there is a check for the state of the definition.

It seems that this might be a cache issue in my cluster environment that has multiple pods. Can someone confirm if this is a possible reason for that behavior?

Hi,

I do not think so. My understanding, subscribed entities are fetched directly from the DB and added to the cache after that.

It would be the best to reproduce the issue or check the logs.

Regards

Martin