elasticsearch - Elasticsearch服务器超时

Question

我正在8核/ 16G RAM Qbox服务器上运行ES节点。
我正在做一些索引和搜索操作。索引(最大5 /秒)，搜索(最大1e4 /秒)。我的索引大约有2M记录了1.2G的数据。搜索查询大约需要300毫秒的中位数。
我对服务器为何无法处理如此低的流量感到困惑。
贝娄是我的映射，也是我要发送的查询:
对应:

{
"orders_prod-1586935194034930": {
    "mappings": {
    "_doc": {
        "dynamic": "true",
        "properties": {
        "address_id": {
            "type": "long"
        },
        "card_4_digits": {
            "type": "keyword"
        },
        "card_type": {
            "type": "keyword"
        },
        "channel_id": {
            "type": "integer"
        },
        "created_at": {
            "type": "date",
            "format": "strict_date_hour_minute_second"
        },
        "customer_email": {
            "type": "keyword"
        },
        "customer_name": {
            "type": "keyword"
        },
        "customer_name_text": {
            "type": "text"
        },
        "customer_phone": {
            "type": "keyword"
        },
        "delivery_address": {
            "type": "keyword"
        },
        "favorite": {
            "type": "boolean"
        },
        "has_promo_discount": {
            "type": "boolean"
        },
        "id": {
            "type": "long"
        },
        "ivr_code": {
            "type": "keyword"
        },
        "order_at": {
            "type": "date",
            "format": "strict_date_hour_minute_second"
        },
        "order_type": {
            "type": "keyword"
        },
        "payment_type": {
            "type": "keyword"
        },
        "placed_at": {
            "type": "date",
            "format": "strict_date_hour_minute_second"
        },
        "restaurant_id": {
            "type": "integer"
        },
        "short_uuid": {
            "type": "keyword"
        },
        "state": {
            "type": "keyword"
        },
        "subtotal": {
            "type": "integer"
        },
        "total": {
            "type": "integer"
        },
        "updated_at": {
            "type": "date",
            "format": "strict_date_hour_minute_second"
        },
        "user_id": {
            "type": "integer"
        },
        "uuid": {
            "type": "keyword"
        },
        "workflow_color": {
            "type": "keyword"
        },
        "workflow_node_id": {
            "type": "keyword"
        },
        "workflow_node_name": {
            "type": "keyword"
        },
        "workflow_tag": {
            "type": "keyword"
        }
        }
    }
    }
}
}

查询:

{
"from":0,
"size":20,
"query":{
    "bool":{
        "must":[
            {
            "bool":{
                "should":[
                    {
                        "term":{
                        "user_id":{
                            "value":308612,
                            "boost":1.0
                        }
                        }
                    },
                    {
                        "term":{
                        "restaurant_id":{
                            "value":8898,
                            "boost":1.0
                        }
                        }
                    },
                    {
                        "term":{
                        "restaurant_id":{
                            "value":4164,
                            "boost":1.0
                        }
                        }
                    },
                    {
                        "term":{
                        "restaurant_id":{
                            "value":4679,
                            "boost":1.0
                        }
                        }
                    }
                ],
                "adjust_pure_negative":true,
                "minimum_should_match":"1",
                "boost":1.0
            }
            }
        ],
        "must_not":[
            {
            "term":{
                "state":{
                    "value":"created",
                    "boost":1.0
                }
            }
            }
        ],
        "should":[
            {
            "bool":{
                "should":[
                    {
                        "term":{
                        "state":{
                            "value":"executing",
                            "boost":1.0
                        }
                        }
                    },
                    {
                        "term":{
                        "state":{
                            "value":"missed",
                            "boost":1.0
                        }
                        }
                    }
                ],
                "adjust_pure_negative":true,
                "minimum_should_match":"1",
                "boost":1.0
            }
            }
        ],
        "adjust_pure_negative":true,
        "minimum_should_match":"1",
        "boost":1.0
    }
},
"_source":{
    "includes":[
        "short_uuid",
        "uuid",
        "state",
        "restaurant_id",
        "channel_id",
        "customer_name_text",
        "total",
        "payment_type",
        "card_type",
        "order_at"
    ],
    "excludes":[

    ]
},
"sort":[
    {
        "order_at":{
            "order":"asc"
        }
    },
    {
        "created_at":{
            "order":"desc"
        }
    }
]
}

有什么我可以做的吗？也许查询根本不是最优的。我真的需要一些提示。
另一个信息。我在ES日志中看到的崩溃:

[2020-08-19T12:03:16,912][DEBUG][o.e.a.s.TransportSearchAction] [s-f8cb43563fe76e4a-1] [orders_prod-1586935194034930][3], node[xkui_jc_TZ20-IJJ4bo1TQ], [P], s[STARTED], a[id=vmh_Z8BHTsGpLoUoseuKYw]: Failed to execute [SearchRequest{searchType=QUERY_THEN_FETCH, indices=[orders_prod], indicesOptions=IndicesOptions[ignore_unavailable=false, allow_no_indices=true, expand_wildcards_open=true, expand_wildcards_closed=false, allow_aliases_to_multiple_indices=true, forbid_closed_indices=true, ignore_aliases=false], types=[_doc], routing='null', preference='null', requestCache=null, scroll=null, maxConcurrentShardRequests=10, batchedReduceSize=512, preFilterShardSize=128, allowPartialSearchResults=true, source={"from":0,"size":20,"query":{"bool":{"must":[{"bool":{"should":[{"term":{"user_id":{"value":18048,"boost":1.0}}},{"term":{"restaurant_id":{"value":800,"boost":1.0}}}],"adjust_pure_negative":true,"minimum_should_match":"1","boost":1.0}},{"term":{"state":{"value":"canceled","boost":1.0}}},{"range":{"order_at":{"from":"2020-08-19T12:02:09","to":"2020-09-19T12:02:09","include_lower":true,"include_upper":true,"boost":1.0}}}],"must_not":[{"term":{"state":{"value":"created","boost":1.0}}}],"adjust_pure_negative":true,"boost":1.0}},"_source":{"includes":["short_uuid","uuid","state","restaurant_id","channel_id","customer_name_text","total","payment_type","card_type","order_at"],"excludes":[]},"sort":[{"order_at":{"order":"asc"}},{"created_at":{"order":"desc"}}]}}] lastShard [true]
org.elasticsearch.transport.RemoteTransportException: [s-f8cb43563fe76e4a-0][10.244.23.158:9300][indices:data/read/search[phase/query]]
Caused by: org.elasticsearch.common.util.concurrent.EsRejectedExecutionException: rejected execution of org.elasticsearch.common.util.concurrent.TimedRunnable@7ea74209 on QueueResizingEsThreadPoolExecutor[name = s-f8cb43563fe76e4a-0/search, queue capacity = 1000, min queue capacity = 1000, max queue capacity = 1000, frame size = 2000, targeted response rate = 1s, task execution EWMA = 625.6ms, adjustment amount = 50, org.elasticsearch.common.util.concurrent.QueueResizingEsThreadPoolExecutor@26e28e[Running, pool size = 7, active threads = 7, queued tasks = 1000, completed tasks = 144056]]
    at org.elasticsearch.common.util.concurrent.EsAbortPolicy.rejectedExecution(EsAbortPolicy.java:48) ~[elasticsearch-6.4.3.jar:6.4.3]
    at java.util.concurrent.ThreadPoolExecutor.reject(ThreadPoolExecutor.java:830) ~[?:1.8.0_201]
    at java.util.concurrent.ThreadPoolExecutor.execute(ThreadPoolExecutor.java:1379) ~[?:1.8.0_201]
    at org.elasticsearch.common.util.concurrent.EsThreadPoolExecutor.doExecute(EsThreadPoolExecutor.java:98) ~[elasticsearch-6.4.3.jar:6.4.3]

Answer 1

你说:

Search query takes around 300ms median. I am puzzled on why the servercan not handle even such a low traffic.

我想您觉得300ms太长了。这并不意味着服务器不处理流量，只是查询有点慢。
首先，您的索引有多少个分片。分片大小的目标应为50Gb左右。因此，如果此1.2GB索引具有多个分片和1个副本，那就太多了。
如果可以为此索引计划维护，则应使用适当数量的分片重新索引并强制合并。
关于您的映射，没有很多更改要做。
通常，ID仅用于严格的相等性，因此我们过去将其映射为关键字，而不是整数(user_id，id，channel_id，address_id等)，但实际上不会对perfs产生敏感影响
如果您使用的是最近的Elasticsearch，而不使用分数，则在重新编制索引时，可以考虑对order_at和created_at字段进行索引排序。它应该有助于节省时间。
https://www.elastic.co/guide/en/elasticsearch/reference/master/index-modules-index-sorting.html
在查询本身上:
您应该考虑使用过滤器。
过滤器不影响得分，并且可在 flex 和OS端进行缓存。
在您的用例中，整个索引可以保留在RAM中。这可以带来巨大的变化。
您应该具有类似以下的行为(必须经过测试):

{
  "size": 20,
  "query": {
    "bool": {
      "filter": {
        "bool": {
          "should": [
            {
              "term": {
                "restaurant_id": "8898"
              }
            },
            {
              "term": {
                "restaurant_id": "4164"
              }
            },
            {
              "term": {
                "restaurant_id": "4679"
              }
            },
            {
              "term": {
                "user_id": "308612"
              }
            },
            {
              "term": {
                "state": "executing"
              }
            },
            {
              "term": {
                "state": "missed"
              }
            }
          ],
          "minimum_should_match": "1",
          "must_not": [
            {
              "term": {
                "state": "created"
              }
            }
          ]
        }
      }
    }
  },
  "_source": [
    "short_uuid",
    "uuid",
    "state",
    "restaurant_id",
    "channel_id",
    "customer_name_text",
    "total",
    "payment_type",
    "card_type",
    "order_at"
  ],
  "sort": [
    {
      "order_at": {
        "order": "asc"
      }
    },
    {
      "created_at": {
        "order": "desc"
      }
    }
  ]
}

如果不使用_score，则可以使用固定分数查询来避免 https://www.elastic.co/guide/en/elasticsearch/reference/current/query-dsl-constant-score-query.html的开销