我想看看一天中按小时划分的直方图。
问题出在 * 时区 *：如果我把一个文档放在美国的上午8点和欧洲的上午8点，我有不同的时区，但在图表中，我想把它们聚集在同一个小时。
因此，我的结果应该是上午8点2个文档计数
我的实际数据：

[
  {
    "_index" : "test_index",
    "_type" : "_doc",
    "_id" : "1",
    "_score" : 1.0,
    "_source" : {
      "my_time" : "2021-09-17T08:00:00.000-04:00",
      "value" : 1
    }
  },
  {
    "_index" : "test_index",
    "_type" : "_doc",
    "_id" : "2",
    "_score" : 1.0,
    "_source" : {
      "my_time" : "2021-09-17T08:00:00.000+02:00",
      "value" : 2
    }
  },
  {
    "_index" : "test_index",
    "_type" : "_doc",
    "_id" : "3",
    "_score" : 1.0,
    "_source" : {
      "my_time" : "2021-09-17T03:00:00.000-04:00",
      "value" : 3
    }
  },
  {
    "_index" : "test_index",
    "_type" : "_doc",
    "_id" : "4",
    "_score" : 1.0,
    "_source" : {
      "my_time" : "2021-09-17T04:00:00.000-08:00",
      "value" : 4
    }
  },
  {
    "_index" : "test_index",
    "_type" : "_doc",
    "_id" : "5",
    "_score" : 1.0,
    "_source" : {
      "my_time" : "2021-09-17T23:00:00.000+04:00",
      "value" : 5
    }
  }
]

在Map中

"my_time": {
    "type": "date"
 }

我的疑问：

POST /test_index/_search?size=0
{
  "aggs": {
    "my_results": {
      "date_histogram": {
        "field": "my_time",
        "calendar_interval": "1h"
      }
    }
  }
}

但是结果是令人担忧的！2！3例如：

"buckets" : [
    {
      "key_as_string" : "2021-09-17T06:00:00.000Z",
      "key" : 1631858400000,
      "doc_count" : 1
    },
    {
      "key_as_string" : "2021-09-17T07:00:00.000Z",
      "key" : 1631862000000,
      "doc_count" : 1
    },
    ......
]

我该如何修复它？