Java实现elastic中服务接口性能指标统计(接口QPS、接口99响应时间等）并存入表

近期需要对线上A服务接口进行健康度监控，即把A服务各个接口每天的性能指标进行统计并写入库表，便于对接口通过周期性数据进行全面分析及接口优化。

据调研了解，有2种解决方案可行：

1）目前服务接口请求信息已全面接入elk(Elasticsearch,Logstash,Kibana)，可以统计kibana中的visualization 功能统计，通过程序模拟visualization功能请求，获取到性能指标数据。

优点：不需要程序进行任何改动，无研发同事时间成本。

缺点：需要进行尝试，未使用此种方式统计过。

2）研发同事在程序中，使用prometheus，对各接口调用时进行埋点，采集接口请求、返回信息。

Prometheus介绍 - 运维人在路上 - 博客园

优点：prometheus采集数据日常较常用、成熟，且采集后可以进行多维度统计数据。

缺点：需要研发同事介入、增加了研发时间成本。

经过考虑，先尝试方案一（后经实践，可行）。

以下为方案一的具体实现细节，供参考。

一、在elastic中，确定好服务查询索引及查询条件。

    查询索引：logstash-lb-nginx-logs-xxxx-*

    查询条件：http_host: "d-es-xxxx.xxxx.com" AND request_uri: "/kw/segmentWordList"

二、使用kibana的visualization功能，画出接口99响应时间及QPS统计图。

最终展示：接口99响应时间及QPS统计图详细创建步骤请参见：elastic中创建服务接口99响应时间及QPS统计图_见贤思齐IT的博客-CSDN博客，在这就不赘述。

接口99分位响应时间，如下图所示：

接口QPS，如下图所示：

三、通过展示图，获取请求body信息（json)。

以下为接口qps为示例，详细介绍下步骤：

1、在dashboards中查询已创建的可视面板。

2、打开inspect

3、在弹出的页面中，选择Request选项

4、终于看到心心念的json请求信息了。

json信息如下：

{
"aggs": {
"2": {
"date_histogram": {
"field": "@timestamp",
"fixed_interval": "1s",
"time_zone": "Asia/Shanghai",
"min_doc_count": 1
},
"aggs": {
"1": {
"percentiles": {
"field": "request_time",
"percents": [
95,
99
],
"keyed": false
}
}
}
}
},
"size": 0,
"fields": [
{
"field": "@timestamp",
"format": "date_time"
},
{
"field": "kafka.block_timestamp",
"format": "date_time"
}
],
"script_fields": {},
"stored_fields": [
"*"
],
"runtime_mappings": {},
"_source": {
"excludes": []
},
"query": {
"bool": {
"must": [],
"filter": [
{
"bool": {
"filter": [
{
"bool": {
"should": [
{
"match_phrase": {
"http_host": "d-es-xx.zpidc.com"
}
}
],
"minimum_should_match": 1
}
},
{
"bool": {
"should": [
{
"match_phrase": {
"request_uri": "/kw/segmentWordList"
}
}
],
"minimum_should_match": 1
}
}
]
}
},
{
"range": {
"@timestamp": {
"gte": "2022-03-30T07:29:17.066Z",
"lte": "2022-03-31T07:29:17.066Z",
"format": "strict_date_optional_time"
}
}
}
],
"should": [],
"must_not": []
}
}
}

四、以下为java代码具体实现，仅供参考。

注：第三节仅是单个接口的展示，在java代码中可以获取服务下所有接口，并参数化请求信息。

package statistical;

import com.alibaba.fastjson.JSON;
import com.alibaba.fastjson.JSONArray;
import com.alibaba.fastjson.JSONObject;
import com.zhilian.statistics.utils.HttpGetPost;
import org.jooq.DSLContext;
import org.jooq.Record;
import org.jooq.Result;
import org.jooq.SQLDialect;
import org.jooq.impl.DSL;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import org.springframework.util.StringUtils;

import java.sql.Connection;
import java.sql.DriverManager;
import java.sql.SQLException;
import java.time.LocalDate;
import java.util.ArrayList;
import java.util.Date;
import java.util.List;

import static com.zhilian.statistics.jooq.public_.Tables.*;

/**
 * @Author: jiajun.du
 * @Date: 2022/3/30 20:50
 */
public class StatisticsNlp {
    private static Logger logger = LoggerFactory.getLogger(StatisticsNlp.class);
    HttpGetPost httpGetPost = new HttpGetPost();
    //连接数据库配置信息
    public static String userName = "xxxx";
    public static String password = "xxxx";
    public static String url = "jdbc:postgresql://172.xx.xx.xx:1601/business_data_platform?tcpKeepAlive=true&autoReconnect=true";
    public static Connection conn;
    public static DSLContext dsl;
    private static long day = 0;
    public static LocalDate localDate = LocalDate.now().plusDays(day);//获取当前日期
    static {
        try {
            conn = DriverManager.getConnection(url, userName, password); //创建连接
            dsl = DSL.using(conn, SQLDialect.POSTGRES);//将连接和具体的数据库类型绑定
        } catch (SQLException e) {
            e.printStackTrace();
        }
    }

    //计算服务接口最大值（99分位耗时、最大QPS）
    public float maxValue(JSONArray array,String name){
        float max99Value = 0f;
        float param = 0f;
        String valueStr = null;
        if(array!=null && array.size()>=1){
            for (int k = 0; k < array.size(); k++) {
                if(name.equals("time_99line")){//99分位耗时统计
                    valueStr = array.getJSONObject(k).getJSONObject("1").getString("values");
                    String p  = valueStr.replace("[", "").replace("]", "");
                    JSONObject jsonObject = JSONObject.parseObject(p);
                    param = Float.parseFloat(jsonObject.getString("value"));
                }else if(name.equals("max_pqs")){//最大QPS统计
                    param = Float.parseFloat(array.getJSONObject(k).getString("doc_count"));
                }
                if (max99Value < param) {
                    max99Value = param;
                }
            }
        }
        return max99Value;
    }

    //更新服务字典表中接口列表信息
    public Boolean nlpMethodUpdate(String index,String body)throws Exception{
        JSONObject methodJsonResult = null;//存放请求结果json
        JSONObject bodyJson = null;//查询body信息
        JSONArray methodJsonArray = null;
        String strParma = null; //临时方法名称变量
        String realmName = "d-es-thirdparty.zpidc.com";
        String moudle = null;
        String strs [] = null;

        List<String [][]> nlpMethods = new ArrayList<>();

        if(url!=null){
            bodyJson = JSONObject.parseObject(body);
            methodJsonResult = HttpGetPost.doPost(index, bodyJson, "UTF-8");
            if(methodJsonResult!=null && methodJsonResult.getJSONObject("aggregations").size()>=1){
                methodJsonArray = methodJsonResult.getJSONObject("aggregations").getJSONObject("2").getJSONArray("buckets");//获取请求结果中array信息
                for(int i = 0;i<methodJsonArray.size();i++){
                    String method [][] = new String[1][2];
                    strParma = methodJsonArray.getJSONObject(i).getString("key");//获取接口方法名称
                    if(strParma!=null && strParma.length()>1){
                        strParma = strParma.replace("http://"+realmName, "");
                        strs = strParma.split("/");
                        moudle = strs[1];
                        method[0][0] = strParma;
                        method[0][1] = moudle;
                        nlpMethods.add(i,method);
                    }
                }
            }
            if(nlpMethods!=null && nlpMethods.size()>=1){
                for(String [][] methodName : nlpMethods){
                    List<Record> r = dsl.select().from(SERVICE_NLP_DATA_DICTIONARY).where(SERVICE_NLP_DATA_DICTIONARY.METHOD_NAME.equal(methodName[0][0])).fetch();
                    if(r.size()==0){
                        dsl.insertInto(SERVICE_NLP_DATA_DICTIONARY)
                                .set(SERVICE_NLP_DATA_DICTIONARY.SERVICE_NAME,realmName)
                                .set(SERVICE_NLP_DATA_DICTIONARY.METHOD_NAME,methodName[0][0])
                                .set(SERVICE_NLP_DATA_DICTIONARY.REMARK,"update")
                                .set(SERVICE_NLP_DATA_DICTIONARY.MODULE,moudle)
                                .returning(SERVICE_NLP_DATA_DICTIONARY.ID)
                                .fetchOne();
                    }
                }
            }
        }
        return null;
    }
    //统计nlp服务接口99分位耗时、maxqps性能指标
    public Boolean nlpDataSource(String index,String [][] querys,Record r) throws Exception {
        JSONObject jsonObjectTimeBody = null;
        JSONObject jsonObjectQpsBody =null;
        JSONObject jsonObjectTimeResponse = null;
        JSONObject jsonObjectQpsResponse = null;
        float time_99 = 0f;//99分位耗时
        float max_qps =0f;
        if(querys!=null&&querys.length>=1){
            JSONArray jsonArrayTime = null;
            JSONArray jsonArrayQps = null;
            for(int i=0;i< querys.length;i++) {
                if (querys[i][0].equals("time_99line")) {
                    jsonObjectTimeBody = JSON.parseObject(querys[i][1]);
                    jsonObjectTimeResponse = HttpGetPost.doPost(index, jsonObjectTimeBody, "UTF-8");
                } else if (querys[i][0].equals("max_pqs")) {
                    jsonObjectQpsBody = JSON.parseObject(querys[i][1]);
                    jsonObjectQpsResponse = HttpGetPost.doPost(index, jsonObjectQpsBody, "UTF-8");
                }
            }
            if(jsonObjectTimeResponse!=null && jsonObjectTimeResponse.getJSONObject("aggregations").size()>=1){
                jsonArrayTime = jsonObjectTimeResponse.getJSONObject("aggregations").getJSONObject("2").getJSONArray("buckets");
                time_99 = maxValue(jsonArrayTime,"time_99line");
            }
            if(jsonObjectQpsResponse!=null && jsonObjectQpsResponse.getJSONObject("aggregations").size()>=1){
                jsonArrayQps = jsonObjectQpsResponse.getJSONObject("aggregations").getJSONObject("2").getJSONArray("buckets");
                max_qps = maxValue(jsonArrayQps,"max_pqs");
            }
            //服务接口性能指标信息入库
            dsl.insertInto(SERVICE_NLP_INTERFACE_INDICATORS)
                    .set(SERVICE_NLP_INTERFACE_INDICATORS.SERVICE_NAME,r.getValue("service_name").toString())
                    .set(SERVICE_NLP_INTERFACE_INDICATORS.MODULE,r.getValue("module").toString())
                    .set(SERVICE_NLP_INTERFACE_INDICATORS.METHOD_NAME,r.getValue("method_name").toString())
                    .set(SERVICE_NLP_INTERFACE_INDICATORS.MAX_99_LINE,time_99)
                    .set(SERVICE_NLP_INTERFACE_INDICATORS.MAX_QPS,max_qps)
                    .set(SERVICE_NLP_INTERFACE_INDICATORS.DATE,localDate.plusDays(day))
                    .returning(SERVICE_NLP_INTERFACE_INDICATORS.ID)
                    .fetchOne();
        }
        return null;
    }

    public static void main(String[] args) throws Exception {
        StatisticsNlp statisticsNlp = new StatisticsNlp();
        String index = "http://ops-xx-xx.xx.com/logstash-lb-nginx-logs-xxxx-*/_search";//查询索引;//接口、渠道级别请求地址前缀
        //统计线上A服务下所有接口信息（body体信息）
        String bodyStr = "{\"aggs\":{\"2\":{\"terms\":{\"field\":\"url_path.keyword\",\"order\":{\"_count\":\"desc\"},\"size\":100}}},\"size\":0,\"fields\":[{\"field\":\"@timestamp\",\"format\":\"date_time\"},{\"field\":\"kafka.block_timestamp\",\"format\":\"date_time\"}],\"script_fields\":{},\"stored_fields\":[\"*\"],\"runtime_mappings\":{},\"_source\":{\"excludes\":[]},\"query\":{\"bool\":{\"must\":[],\"filter\":[{\"bool\":{\"should\":[{\"match_phrase\":{\"http_host\":\"d-es-xxxx.xxxx.com\"}}],\"minimum_should_match\":1}},{\"range\":{\"@timestamp\":{\"gte\":\""+localDate.plusDays(day-3)+"T00:00:00.000Z\",\"lte\":\""+localDate.plusDays(day)+"T23:59:00.000Z\",\"format\":\"strict_date_optional_time\"}}}],\"should\":[],\"must_not\":[]}}}";
        statisticsNlp.nlpMethodUpdate(index,bodyStr);//更新服务接口列表
        Result<Record> nlp_result = dsl.select().from(SERVICE_NLP_DATA_DICTIONARY).fetch();//获取服务列表
        String [][] querys = new String[2][2];//存放请求信息
        for(int t = 0 ; t < 1; t++){
            day = t * -1;
            for(Record r:nlp_result){//遍历nlp服务接口
                //time_99line
                querys[0][0] = "time_99line";
                querys[0][1] = "{\"aggs\":{\"2\":{\"date_histogram\":{\"field\":\"@timestamp\",\"fixed_interval\":\"1m\",\"time_zone\":\"Asia/Shanghai\",\"min_doc_count\":1},\"aggs\":{\"1\":{\"percentiles\":{\"field\":\"request_time\",\"percents\":[99],\"keyed\":false}}}}},\"size\":0,\"fields\":[{\"field\":\"@timestamp\",\"format\":\"date_time\"},{\"field\":\"kafka.block_timestamp\",\"format\":\"date_time\"}],\"script_fields\":{},\"stored_fields\":[\"*\"],\"runtime_mappings\":{},\"_source\":{\"excludes\":[]},\"query\":{\"bool\":{\"must\":[],\"filter\":[{\"bool\":{\"filter\":[{\"bool\":{\"should\":[{\"match_phrase\":{\"http_host\":\"d-es-xxxx.xxxx.com\"}}],\"minimum_should_match\":1}},{\"bool\":{\"should\":[{\"match_phrase\":{\"request_uri\":\""+r.getValue("method_name")+"\"}}],\"minimum_should_match\":1}}]}},{\"range\":{\"@timestamp\":{\"gte\":\""+localDate.plusDays(day)+"T08:00:00.066Z\",\"lte\":\""+localDate.plusDays(day)+"T23:00:17.066Z\",\"format\":\"strict_date_optional_time\"}}}],\"should\":[],\"must_not\":[]}}}";
                //maxqps
                querys[1][0] = "max_pqs";
                querys[1][1] = "{\"aggs\":{\"2\":{\"date_histogram\":{\"field\":\"@timestamp\",\"fixed_interval\":\"1s\",\"time_zone\":\"Asia/Shanghai\",\"min_doc_count\":1}}},\"size\":0,\"fields\":[{\"field\":\"@timestamp\",\"format\":\"date_time\"},{\"field\":\"kafka.block_timestamp\",\"format\":\"date_time\"}],\"script_fields\":{},\"stored_fields\":[\"*\"],\"runtime_mappings\":{},\"_source\":{\"excludes\":[]},\"query\":{\"bool\":{\"must\":[],\"filter\":[{\"match_all\":{}},{\"bool\":{\"filter\":[{\"bool\":{\"should\":[{\"match_phrase\":{\"http_host\":\"d-es-xxxx.xxxx.com\"}}],\"minimum_should_match\":1}},{\"bool\":{\"should\":[{\"match_phrase\":{\"request_uri\":\""+r.getValue("method_name")+"\"}}],\"minimum_should_match\":1}}]}},{\"range\":{\"@timestamp\":{\"gte\":\""+localDate.plusDays(day)+"T08:00:09.025Z\",\"lte\":\""+localDate.plusDays(day)+"T23:00:12.479Z\",\"format\":\"strict_date_optional_time\"}}}],\"should\":[],\"must_not\":[]}}}";
                statisticsNlp.nlpDataSource(index,querys,r);
            }
        }
    }
}

package com.zhilian.statistics.utils;

import com.alibaba.fastjson.JSONObject;
import org.apache.http.HttpEntity;
import org.apache.http.client.ClientProtocolException;
import org.apache.http.client.config.RequestConfig;
import org.apache.http.client.methods.CloseableHttpResponse;
import org.apache.http.client.methods.HttpPost;
import org.apache.http.entity.StringEntity;
import org.apache.http.impl.client.CloseableHttpClient;
import org.apache.http.impl.client.HttpClients;
import org.apache.http.util.EntityUtils;a
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;

import java.io.IOException;
import java.text.ParseException;

/**
 * @Author: jiajun.du
 * @Date: 2021/11/23 15:53
 */
public class HttpGetPost {
    Logger logger = LoggerFactory.getLogger(HttpGetPost.class);
    /**
     * get请求
     * @param url
     * @return
     * @throws Exception
     */
     public JSONObject doGet(String url, String param) throws Exception{
        CloseableHttpClient httpClient = null;
        CloseableHttpResponse httpResponse =null;
        String strResult = "";
        try {
            httpClient = HttpClients.createDefault();//创建一个httpClient实例
            org.apache.http.client.methods.HttpGet httpGet =new org.apache.http.client.methods.HttpGet(url+param);//创建httpGet远程连接实例
            // 设置请求头信息
            RequestConfig requestConfig = RequestConfig.custom().setConnectTimeout(300000).//连接主机服务器时间
                    setConnectionRequestTimeout(300000).//请求超时时间
                    setSocketTimeout(300000).//数据读取超时时间
                    build();
            httpGet.setConfig(requestConfig);//为httpGet实例设置配置信息
            httpResponse = httpClient.execute(httpGet);//通过get请求得到返回对象

            //发送请求成功并得到响应
            if(httpResponse.getStatusLine().getStatusCode()==200){
                strResult = EntityUtils.toString(httpResponse.getEntity());
                JSONObject resultJsonObject = JSONObject.parseObject(strResult);//获取请求返回结果
                return resultJsonObject;
            }else{
                logger.error("请求{}失败,\n状态码为{}",url,httpResponse.getStatusLine().getStatusCode());
            }
        } catch (ClientProtocolException e) {
            e.printStackTrace();
        } catch (IOException i){
            i.printStackTrace();
            logger.error("IOException异常：{}",i.getMessage());
        } finally {
            if(httpResponse!=null){
                try {
                    httpResponse.close();
                } catch (IOException e) {
                    e.printStackTrace();
                }
            }
            if(httpClient!=null){
                try {
                    httpClient.close();
                } catch (IOException e) {
                    e.printStackTrace();
                }
            }
        }
        return null;
    }

    /**
     * 发送post请求
     * @param url  路径
     * @param jsonObject  参数(json类型)
     * @param encoding 编码格式
     * @return
     * @throws ParseException
     * @throws IOException
     */
    public static JSONObject doPost(String url, JSONObject jsonObject,String encoding) throws ParseException, IOException{
        JSONObject object = null;
        String body = "";

        //创建httpclient对象
        CloseableHttpClient client = HttpClients.createDefault();
        //创建post方式请求对象
        HttpPost httpPost = new HttpPost(url);

        //装填参数
        StringEntity s = new StringEntity(jsonObject.toString(), "utf-8");
        //s.setContentEncoding(new BasicHeader(HTTP.CONTENT_TYPE, "application/json"));
        //设置参数到请求对象中
        httpPost.setEntity(s);
        System.out.println("请求地址："+url);

        //设置header信息
        //指定报文头【Content-type】、【User-Agent】
        httpPost.setHeader("Content-Type", "application/json");
        httpPost.setHeader("User-Agent", "Mozilla/4.0 (compatible; MSIE 5.0; Windows NT; DigExt)");
        httpPost.setHeader("Connection","keep-alive");
        
        //执行请求操作，并拿到结果（同步阻塞）
        CloseableHttpResponse response = client.execute(httpPost);
        //获取结果实体
        HttpEntity entity = response.getEntity();
        if (entity != null) {
            //按指定编码转换结果实体为String类型
            body = EntityUtils.toString(entity, encoding);
            object = JSONObject.parseObject(body);
        }
        EntityUtils.consume(entity);
        //释放链接
        response.close();
        return object;
    }
}

落库结果一览（隐私信息已模糊处理）：

标签： elasticsearch java 压力测试

本文转载自: https://blog.csdn.net/junior77/article/details/123897416
版权归原作者 ~见贤思齐~ 所有，如有侵权，请联系我们删除。

Java实现elastic中服务接口性能指标统计(接口QPS、接口99响应时间等）并存入表

发表评论

“Java实现elastic中服务接口性能指标统计(接口QPS、接口99响应时间等）并存入表”的评论:

关于作者

overfit同步小助手

相关阅读

文章导航