I have set the properties:
indices.fielddata.cache.size: 65%
indices.breaker.fielddata.limit: 70%
but my old gen also can reach 100%, then frequent full gc . I should restart my ES cluster for the state.
any body know how to handle about this problem ? thank you !