ES7.16.2基础操作之percentiles百分比请求时长（十七）

本文链接：https://blog.csdn.net/ls_call520/article/details/122268184

这篇博客探讨了如何通过百分位数统计方法来分析网站请求的延迟情况。内容包括全国不同省份的请求耗时分布，如tp50、tp90、tp95等指标，以及各省份的平均响应时间。此外，还使用了percentile_ranks统计来了解特定耗时请求所占的百分比，以辅助进行性能调优。

摘要由CSDN通过智能技术生成

## 初始化数据 
POST websit/logs/_bulk
 {"index":{}}
 {"latency":105,"province":"湖南","timestamp":"2021-10-28"}
 {"index":{}}
 {"latency":60,"province":"广东","timestamp":"2021-12-28"}
 {"index":{}}
 {"latency":70,"province":"广西","timestamp":"2021-10-28"}
 {"index":{}}
 {"latency":85,"province":"湖北","timestamp":"2022-01-01"}
 {"index":{}}
 {"latency":90,"province":"四川","timestamp":"2022-01-01"}
 {"index":{}}
 {"latency":100,"province":"湖南","timestamp":"2022-01-01"}
 {"index":{}}
 {"latency":65,"province":"江苏","timestamp":"2022-01-01"}
 {"index":{}}
 {"latency":75,"province":"广东","timestamp":"2021-10-28"}
 {"index":{}}
 {"latency":80,"province":"新疆","timestamp":"2021-12-31"}
 {"index":{}}
 {"latency":50,"province":"湖南","timestamp":"2021-07-31"}
 
## percentiles百分比统计
##tp50：50%的请求耗时最长在多长时间
##tp90：90%的请求耗时最长在多长时间
##tp95：95%的请求耗时最长在多长时间
GET websit/_search
{
  "size": 0,
  "aggs": {
    "latency_percentiles": {
      "percentiles": {
        "field": "latency",
        "percents": [
          50,
          90,
          95
        ]
      }
    },
    "latency_avg": {
      "avg": {
        "field": "latency"
      }
    }
  }
}

## 百分比统计：按省份分别统计请求时长
GET websit/_search
{
  "size": 0,
  "aggs": {
    "group_by_province": {
      "terms": {
        "field": "province.keyword"
      },
      "aggs": {
        "latency_percentiles": {
          "percentiles": {
            "field": "latency",
            "percents": [
              50,
              90,
              95
            ]
          }
        },
        "latency_avg": {
          "avg": {
            "field": "latency"
          }
        }
      }
    }
  }
}

## percentile_ranks统计网站访问请求
##如100ms以内请求占百分比,200ms请求的占百分比
## 性能调优：compression值越大，性能越好
GET websit/_search
{
  "size": 0,
  "aggs": {
    "per_ranks": {
      "terms": {
        "field": "province.keyword"
      },
      "aggs": {
        "load_times": {
          "percentile_ranks": {
            "compression": 100, 
            "field": "latency",
            "values": [
              100,
              200
            ]
          }
        }
      }
    }
  }
}