We setup some monitoring to watch for parent circuit breaker trips in Elasticsearch and its going off like crazy. What I am trying to figure out is how to determine what is causing it. From what I have read and understand typically these go off when there is a risk of low memory for the JVM however that could be because we've written poor queries or it could be because we've undersized our Elasticsearch instances in terms of memory. I am trying to figure out how to determine which is the case and exactly what errors are occurring.
I tried to look at the Elasticsearch logs but I am not seeing anything being logged (the exceptions we're getting are in monitoring - not logging). Is there some way to enable logging of these errors - ideally with the query that triggered the alert?
Thanks
Brad