Document workflow events in CMS get triggered with a delay causing issues with event listener implementation

On one of our Bloomreach cloud environments we’re experiencing a delay in the triggering of our document workflow event listener implementation.

This delayed behaviour becomes problematic when there is a sequence of document workflow that gets triggered consistenly (e.g. save, delete, publish, take offline) because that will result in a queue of events that get triggered one by one and is causing more delay. E.g. a new event has to wait for all existing events to get triggered. In our case, we sometimes have to wait for more than 2 hours! before an event gets triggered. It’s extreme.

This only happens on one of the environments running on the same Bloomreach cloud. On other envrionments events get triggered near instant. (None of the environments is yet in production).

Anyone knows what may cause this delayed behaviour and how to fix / avoid?