apache benchmark 做接口压力测试

1. 安装压测工具包，httpd-tools

 

yum install httpd-tools

 

2. post 请求将参数放在文件里面，(可根据自己项目或者公司里面的接口请求，fiddler抓包获取对应的demo数据和地址)

159_root -->cat GetRsOrderInfo.txt

{"method":"GetRsOrderInfo","orderId":2785629,"version":"v0.1.0","localHSFVersion":"8.2.574","localCKVersion":"9.1.498","nativeVersion":"ljck-ios-5.2.4","token":"5df867578f5740ba98ccb57ef12f1b30"}

 

3. 开始压测

-n：执行的请求个数，默认时执行一个请求

-c：一次产生的请求个数，即并发个数

-p： 模拟 POST 请求，文件格式为 param1=1&param2=2，配合 -T 使用

-T：POST 数据所使用的 Content-Type 头信息，如 -T application/x-www-form-urlencoded

 

159_root -->ab -c 3 -n 100 -p GetRsOrderInfo.txt -T application/x-www-form-urlencoded http://txapi.xxxxx.com/api/GetRsOrderInfo.json

This is ApacheBench, Version 2.3 <$Revision: 1843412 $>

Copyright 1996 Adam Twiss, Zeus Technology Ltd, http://www.zeustech.net/

Licensed to The Apache Software Foundation, http://www.apache.org/

 

Benchmarking txapi.xxxxx.com (be patient).....done

 

 

Server Software: nginx/1.10.2

Server Hostname: txapi.xxxxx.com (域名)

Server Port: 80

 

Document Path: /api/GetRsOrderInfo.json (地址)

Document Length: 15376 bytes

 

Concurrency Level: 3 (并发数)

Time taken for tests: 4.044 seconds (总请求时间)

Complete requests: 100 (完成请求数)

Failed requests: 0 (失败请求数)

Total transferred: 1569300 bytes

Total body sent: 37300

HTML transferred: 1537600 bytes

Requests per second: 24.73 [#/sec] (mean) (平均每秒请求数)

Time per request: 121.321 [ms] (mean) (平均每个请求消耗时间)

Time per request: 40.440 [ms] (mean, across all concurrent requests)

Transfer rate: 378.96 [Kbytes/sec] received

9.01 kb/s sent

387.96 kb/s total (传输速率)

 

Connection Times (ms)

min mean[+/-sd] median max

Connect: 3 4 0.6 4 8

Processing: 70 114 85.2 78 501

Waiting: 69 91 50.9 76 291

Total: 74 118 85.2 82 504

 

Percentage of the requests served within a certain time (ms)

50% 82 (50% 的请求都在 936ms 内完成)

66% 85

75% 92

80% 110

90% 296

95% 303

98% 317

99% 504

100% 504 (longest request)

 

4. 逐渐加压，查看系统各项指标参数

4.1 查看服务器的cpu、内存、网络；

top -- cpu、内存、进程等

free -- 内存

iostat --输入、输出统计

 

lejia_ceshi_13--> free

total used free shared buff/cache available

Mem: 16257492 11803272 710928 5084 3743292 3939836

Swap: 0 0 0

 

lejia_ceshi_13--> iostat 2 10 （每2s输出一次，输出10次）

Linux 4.18.0-147.8.1.el8_1.x86_64 (master-node) 04/06/2021 _x86_64_ (4 CPU)

 

avg-cpu: %user %nice %system %iowait %steal %idle

2.13 0.00 1.33 0.25 0.00 96.28

 

Device tps kB_read/s kB_wrtn/s kB_read kB_wrtn

vda 35.75 7.49 169.69 199108317 4512901812

vdb 2.08 0.96 28.41 25542501 755646956

 

4.2 查看数据库的内存、cpu、死锁、连接数、数据库延迟等各项指标(mysql 为例)

详见： mysql常用性能指标

show processlist;

SELECT * FROM information_schema.INNODB_TRX\G

SELECT * FROM information_schema.INNODB_LOCK_WAITS;

SELECT * FROM information_schema.INNODB_LOCKS;

 

information_schema>select * from PROCESSLIST limit 5;

+----------+----------+--------------------+--------------+---------+------+-------+------+

| ID | USER | HOST | DB | COMMAND | TIME | STATE | INFO |

+----------+----------+--------------------+--------------+---------+------+-------+------+

| 15680598 | lejia008 | 192.168.0.13:46936 | lejia_helper | Sleep | 1128 | | NULL |

| 15686394 | lejia008 | 192.168.0.13:4145 | lejia_helper | Sleep | 530 | | NULL |

| 15682995 | lejia008 | 192.168.0.13:50686 | lejia_helper | Sleep | 756 | | NULL |

| 15160594 | middle | 192.168.0.13:43068 | lejia_middle | Sleep | 0 | | NULL |

| 15687692 | lejia008 | 192.168.0.13:54966 | lejia_helper | Sleep | 330 | | NULL |

+----------+----------+--------------------+--------------+---------+------+-------+------+

5 rows in set (0.00 sec)

 

information_schema>select * from PROCESSLIST where user='';

 

4.3 查看redis等系统性能

info

 

server:服务器运行的环境参数

clients:客户端相关信息

memory：服务器运行内存统计数据

persistence：持久化信息

stats：通用统计数据

Replication：主从复制相关信息

CPU：CPU使用情况

cluster：集群信息

Keypass：键值对统计数量信息

 

4.4 查看具体项目进程运行情况：

项目进程内存、cpu等消耗

项目具体的日志，有无报错，每个请求处理时间，哪块逻辑消耗的时间长短等

 

04-06 13:42:56 458 - W00002428,/api/GetRsOrderInfo.json,0,true,5df867578f5740ba98ccb57ef12f1b30,55

04-06 13:42:56 486 - W00002429,/api/GetRsOrderInfo.json,0,true,5df867578f5740ba98ccb57ef12f1b30,62

04-06 13:42:56 530 - W00002430,/api/GetRsOrderInfo.json,0,true,5df867578f5740ba98ccb57ef12f1b30,60

 

 

附：mysql常用性能指标

 

(1) QPS(每秒Query量)

QPS = Questions(or Queries) / uptime

mysql> show global status like 'Question%';

mysql> show global status like 'uptime';

 

(2) TPS(每秒事务量)

TPS = (Com_commit + Com_rollback) / uptime

mysql > show global status like 'Com_commit';

mysql > show global status like 'Com_rollback';

mysql> show global status like 'uptime';

 

(3)key Buffer 命中率

mysql>show  global  status  like  'key%';

key_buffer_read_hits = (1-key_reads /key_read_requests) * 100%

key_buffer_write_hits = (1-key_writes /key_write_requests) * 100%

 

(4)InnoDB Buffer命中率

mysql> show status like 'innodb_buffer_pool_read%';

innodb_buffer_read_hits = (1 -innodb_buffer_pool_reads / innodb_buffer_pool_read_requests) * 100%

 

(5)Query Cache命中率

mysql> show status like 'Qcache%';

Query_cache_hits = (Qcahce_hits /(Qcache_hits + Qcache_inserts )) * 100%;

 

(6)Table Cache状态量

mysql> show global  status like 'open%';

比较 open_tables  与opend_tables值

 

(7)Thread Cache 命中率

mysql> show global status like'Thread%';

mysql> show global status like'Connections';

Thread_cache_hits = (1 - Threads_created /connections ) * 100%

 

(8)锁定状态

mysql> show global  status like '%lock%';

Table_locks_waited/Table_locks_immediate=0.3%  如果这个比值比较大的话，说明表锁造成的阻塞比较严重

Innodb_row_lock_waits innodb行锁，太大可能是间隙锁造成的

 

(9)复制延时量

mysql > show slave status

查看Seconds_Behind_Master的值，如果为0，说明没有延迟

 

(10) Tmp Table 状况(临时表状况)

mysql > show global status like 'Created_tmp%';

Created_tmp_disk_tables/Created_tmp_tables比值最好不要超过10%，如果Created_tmp_tables值比较大，

可能是排序句子过多或者是连接句子不够优化

 

(11) Binlog Cache 使用状况

mysql > show global status like 'Binlog_cache%';

如果Binlog_cache_disk_use值不为0 ，可能需要调大 binlog_cache_size大小

 

(12) Innodb_log_waits 量

mysql > show status like'innodb_log_waits';

Innodb_log_waits值不等于0的话，表明 innodblog  buffer 因为空间不足而等待

 

(13)open file and table

mysql> show global status like 'Open_files';

mysql> show global status like 'Open_tables';

 

(14) 慢查询 

开启慢查询：

1).手动执行命令开启:

mysql> set global slow_query_log=on;

mysql> set global long_query_time=1;

 

2).编辑/etc/my.cnf，在[mysqld]域中添加：

slow_query_log= 1   # 开启慢查询

slow_query_log_file=/data/data/localhost-slow.log   # 慢查询日志路径

long_query_time= 1        # 慢查询的时长

 

 

(15)全日志

查看全日志：

  show global variables like 'general_log';

开启全日志：

    set global general_log=on;

注意开启全日志会消耗服务器性能，一般只有在排查问题时才会短暂打开。