Hello. I have a cluster of 3 machines with 16 cores and 32 gigs of RAM. The last 2 weeks there has been a problem that 28-29 GB of 32 GB become full. I sinned on the cache, cleaned it, but the memory did not go away. Rebooting the elastic itself also does not bring results. On 2 machines, another graylog is hanging in parallel. Rebooting it also did not give results. I executed the curl "localhost:9200/_nodes/jvm?pretty&human"
then I give the command log: Output logs , Elasticsearch config, jvm.options vm characteristics:
What other data is needed for diagnostics and understanding what is happening? Elastic is 7.10.2