We are running Elasticsearch 6.3 (Basic License). We gather Indices Stats data, specifially these fields:
I would like to compute Query latency data points throughout the day to gauge what our real-time users are experiencing. Questions:
If I divide the search.query_time_in_millis over the search.query_total will that produce a meaningful Query Latency measurement that I can accurately depict user experience with search?
Does the search.query_time_in_millis represent the total Query Latency, including the Waiting Time, or is this just the Service Time (as described here in this Presentation: https://www.elastic.co/elasticon/conf/2018/sf/the-seven-deadly-sins-of-elasticsearch-benchmarking @12:50)