[ML] Explain Log Rate Spikes: Grouped analysis results improvements. (8.7)

[walterra]

Follow up to #142626

API

• In the clean up step, we remove field/value pairs from groups that are not part of the originally retrieved significant terms. This can result in duplicate item sets if there's already one present without the field/value pair being removed.
  [ML] Explain Log Rate Spikes: Fix item set post processing. #147499
• To track group results internally in the code we use a approach with a record that doesn't support multiple values for the same field
  [ML] Explain Log Rate Spikes: Fix client side code to transform groups into table rows. #147592
• For groups that have multiple values for the same field, the group histogram query isn't able to fetch data because it filters with a bool must and individual term aggregations. In these cases we should group by field and use a terms aggregation if there are multiple values for a field so just these get treated as OR.
  [ML] Explain Log Rate Spikes: Fix mini histograms for groups with multiples values per field. #147597

UI

• For grouped results, indicate in the expanded table row field / value pairs that appear in other groups - see [ML] Explain Log Rate Spikes: improve the display of large groups  #145609 (review)
  • in expanded row - add asterisk with tooltip indicating it's a pair unique to the group
  • [ML] Explain Log Rate Spikes: highlight field pairs unique to groups in expanded row #148601
• Enhance the UX for grouped results which only contain field / value pairs that appear in other groups. Could drop the word more in the label perhaps in this case? (Seen with the kibana-8.3-broken-rule-problem data set):
  - [ML] AIOps: Enhance the UX for grouped results which only contain field / value pairs that appear in other groups #148757

[elasticmachine]

Pinging @elastic/ml-ui (:ml)

[peteharverson]

Closing issue as all planned items have been completed.