We are observing that there are a few nodes in our cluster seem to have a spike in search queue items (correlated with a spike in latency). We are relying on slow logs for some part of debugging and I wasn't sure if the
took_millis logged in slow logs includes the time a request spent in the queue before execution. Is there anyway to get a breakdown of how much time a request spent in the queue?
We are using ES 5.6