Looks like data in serach comes from two sources:
wmf_raw_.cirrussearchrequest:
TestSearchSatisfaction eventlogging schema: https://meta.wikimedia.org/wiki/Schema:TestSearchSatisfaction
The schema data is going into MySQL. The team has another schema (different name, same data) that is used to harvest data for AB test that is persisted only to hadoop so they can send a bigger volume of events:
https://grafana.wikimedia.org/d/000000018/eventlogging-schema?orgId=1&var-schema=SearchSatisfaction