Default precision of cardinality in kibana visualizations

Hi! I'm using kibana 7.5 and i have an issue, in my json files i have an aggregation with features and scenarios by features items, each one have a state (passed or false by feature), and i need to show in a kibana visualization failed and passed features, but it dosen't work, i use the "unique count" by feature.name.keyword and "add a bucket slices" by status.keyword, but it return more features than i really have.