Our clients encountered errors below recently when performing index and query requests:
org.elasticsearch.client.ResponseException: org.elasticsearch.client.ResponseException: method [POST], host [http://192.168.12.171:9201], URI [/_bulk], status line [HTTP/1.1 429 Too Many Requests]
{"error":{"root_cause":[{"type":"circuit_breaking_exception","reason":"[parent] Data too large, data for [<http_request>] would be [30299660794/28.2gb], which is larger than the limit of [28991029248/27gb], real usage: [30299434000/28.2gb], new bytes reserved: [226794/221.4kb], usages [eql_sequence=0/0b, model_inference=0/0b, inflight_requests=226794/221.4kb, request=0/0b, fielddata=389538262/371.4mb]","bytes_wanted":30299660794,"bytes_limit":28991029248,"durability":"PERMANENT"}],"type":"circuit_breaking_exception","reason":"[parent] Data too large, data for [<http_request>] would be [30299660794/28.2gb], which is larger than the limit of [28991029248/27gb], real usage: [30299434000/28.2gb], new bytes reserved: [226794/221.4kb], usages [eql_sequence=0/0b, model_inference=0/0b, inflight_requests=226794/221.4kb, request=0/0b, fielddata=389538262/371.4mb]","bytes_wanted":30299660794,"bytes_limit":28991029248,"durability":"PERMANENT"},"status":429}
We checked the nodes stat, here is result:
The image shows that every node has consumed over 20G heap memory, but in the right columns,detailed memory summed is less then 5G, where are the left memories go? How can I know where it is consumed?