如何在ElasticSearch中获取百分位数 [英] How do I get percentiles with count in ElasticSearch

问题描述

我们正尝试在 ElasticSearch 中使用百分点聚合

我们正在使用ElasticSearch监控我们的API性能。我们存储的一组样本文件是

  {name：GET / login，avg count：5} 
 {name：GET / login，avg：1.5，count：3} 
 {name：GET / login，avg ：6.9，count：1} 
 {name：GET / login，avg：3.1，count：1} 
  pre> 
 
 
其中
 
 
 
 
 
  count 是一个 GET / login 的数量到我们服务器的特定时间段。
 
 
  avg 是平均响应时间。
 
 
 
 
 
我们使用以下查询获取百分位数。
  GET / hist / t / _search 
 {
aggs：{
avg_time_percentile：{
百分位数：{
field：avg，
percents：[75，95，99] 
} 
} 
} 
} 
  
我们得到以下回应
 avg_time_percentile：{
values：{
75.0：3.75，
95.0 ：5.549999999999999，
99.0：5.909999999999999 
} 
  
然而，想要的是：
 avg_time_percentile：{
values：{
75.0：2 ，
95.0：4.349999999999998，
99.0：5.670000000000002 
} 
  
错误值的原因是，它正在采取例如 2.2 仅在一个记录中。但是计数是 5 ，它应该被考虑5次。
 
 
 
我如何在ElasticSearch中执行。 / p> 
 
解决方案
而不是使用字段，您可以使用脚本，以便将多个字段合并在一起并运行这些新值的百分位数
  POST / hist / t $ _ 
 $ b {
aggs：{
avg_time_percentile：{
percentiles：{
script：{
 inline：doc.avg.value * doc.count.value
}，
percents：[75，95，99] 
} 
} 
} 
} 
  
 
We are trying to generate Percentiles in ElasticSearch using Percentiles aggregation

We are monitoring our API performance with ElasticSearch. A sample set of documents that we store are
{"name": "GET /login", "avg":2.2, "count": 5}
{"name": "GET /login", "avg":1.5, "count": 3}
{"name": "GET /login", "avg":6.9, "count": 1}
{"name": "GET /login", "avg":3.1, "count": 1}
Where 


count is number of GET /login web requests coming to our server for a specific time period.
avg is average response time.


We uses following query to get percentile.
GET /hist/t/_search
{
  "aggs": {
    "avg_time_percentile": {
      "percentiles": {
        "field": "avg",
         "percents" : [75, 95, 99] 
      }
    }
  }
}
We get following response
"avg_time_percentile": {
  "values": {
    "75.0": 3.75,
    "95.0": 5.549999999999999,
    "99.0": 5.909999999999999
  }
However, what we want is:
"avg_time_percentile": {
  "values": {
    "75.0": 2,
    "95.0": 4.349999999999998,
    "99.0": 5.670000000000002
  }
The reason for the incorrect value is, it is taking for e.g. 2.2 as only in one record. But the count is 5 and it should be considered 5 times.

How do I do it in ElasticSearch.
解决方案
Instead of using field you can use script in order to combine several fields together and run the percentiles on those new values
POST /hist/t/_search
{
  "aggs": {
    "avg_time_percentile": {
      "percentiles": {
         "script": {
            "inline": "doc.avg.value * doc.count.value"
         },
         "percents" : [75, 95, 99] 
      }
    }
  }
}


                        
这篇关于如何在ElasticSearch中获取百分位数的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！