Discover Top values "Calculated from 5,000 sample records."

[stephenb (Stephen Brown)]

Click Visualize, It will take you to Lens
Then adjust the timepicker to the entire time range if that is meaningful
Note Aggregations on large data set can have small inaccuracies with respect to aggregations
Depending on the version in Lens there is an option to select High Accuracy mode.

It would be worth reading this

There is also an excellent discussion here by one of our experts

Elasticsearch aggregation on million or more data

I have an index which can have millions of documents with time. I need to make several aggregation on those huge amount of documents fields. In these documents, I need to count number of unique value of a field which may grow large with number of documents like more than million. I know "cardinality" aggregation. But this doesn't give right count, this gives approximation. Then I tried to used scripted_metric. { "aggs": { "distinct_count": { …