ab

ab is a pressure testing tool provided by apache, which is very convenient to use.

install

If Apache is installed, then ab is installed. If you don't want to install apache, you can install ab in the following way

# ubuntu
sudo apt-get install apache2-utils

# centos
yum -y install httpd-tools 

Pressure measurement

Before pressure testing, there are several options that need to be looked at through ab-help

Options are:
    -n requests     Number of requests to be executed
    -c concurrency  Concurrent number
    -s timeout      response time

implement

ab -n 1000 -c 100 -s 1 http://127.0.0.1:1080/event?config_name=mysql_config

# A total of 100 requests were executed concurrently with a timeout of 1 s

Result analysis

After executing the above test commands, the following results are obtained, focusing on the following indicators:

• Failed requests: Failed requests
• Requests per second: also known as QPS, query rate per second, which is an average.
• Time per request: The time it takes to complete a request
• Transfer rate: Network transmission speed. For large file request testing, this value can easily become the bottleneck of the system. To determine whether the value is a bottleneck, we need to know the network situation between the client and the server under test, including network bandwidth and network card speed.

Server Software:        
Server Hostname:        127.0.0.1
Server Port:            1080

Document Path:          /event?config_name=mysql_config
Document Length:        0 bytes

Concurrency Level:      100
Time taken for tests:   0.137 seconds
Complete requests:      1000
Failed requests:        0
Total transferred:      75000 bytes
HTML transferred:       0 bytes
Requests per second:    7275.11 [#/sec] (mean)
Time per request:       13.745 [ms] (mean)
Time per request:       0.137 [ms] (mean, across all concurrent requests)
Transfer rate:          532.84 [Kbytes/sec] received

Connection Times (ms)
                min  mean[+/-sd] median   max
Connect:        0    1   1.0      1       5
Processing:     1   12   5.9     11      30
Waiting:        1   11   5.9     11      29
Total:          1   13   6.1     13      30

Percentage of the requests served within a certain time (ms)
  50%     13
  66%     15
  75%     17
  80%     18
  90%     22
  95%     25
  98%     28
  99%     29
 100%     30 (longest request)

 

 

 

 

 

wrk

wrk is a modern http performance benchmarking tool developed in c language, which is simple to use and powerful.

install

Installation can be downloaded from github source code compilation and installation wrk github address

Pressure measurement

wrk has few command options and is easy to use

Usage: wrk <options> <url>                            
  Options:                                            
    -c, --connections <N>  Established connections  
    -d, --duration    <T>  Execution test time           
    -t, --threads     <N>  Number of threads   
                                                      
    -s, --script      <S>  Use lua Script(Very powerful functions, interested in in in-depth study)       
    -H, --header      <H>  For every one HTTP Request add HTTP head        
        --latency          At the end of the pressure measurement, print the delay statistics   
        --timeout     <T>  timeout    
    -v, --version          Print version details      
                                                      
  Numeric arguments may include a SI unit (1k, 1M, 1G)
  Time arguments may include a time unit (2s, 2m, 2h)

wrk -c100 -t10 -d30s http://127.0.0.1:1080/event?config_name=mysql_config

# Ten threads 100 connections manometer 30s

Result analysis

The results report generated by wrk is relatively concise, and the main focus is similar to ab.

• Requests/sec: QPS
• Transfer/sec: Network Transmission Speed

Running 30s test @ http://127.0.0.1:1080/event?config_name=mysql_config
  10 threads and 100 connections
  Thread Stats   Avg      Stdev     Max   +/- Stdev
    Latency    21.02ms   92.04ms   1.05s    97.79%
    Req/Sec     1.22k   229.72     2.38k    75.47%
  365483 requests in 30.07s, 26.14MB read
Requests/sec:  12152.63
Transfer/sec:      0.87MB

 

 

 

jmeter

jmeter is a GUI testing tool written by java. It has powerful functions and diverse results.

install

Installation can go apache jmeter download

Use

Building Plan

Establish Thread Group

Thread Group can be set up after the Plan s have been set up. The way to set up Thread Group

When Plan s are selected, Edit > Add > Threads > Thread Group

Thread Group can set some parameters of threads, mainly Number of Threads(users) and Loop Count.

Add Listener

Listener mainly generates some result reports, which are added as follows

Edit > Add > Listener after Thread Group is selected

You can see a lot of results reports. I usually use the following results reports.

• View Results Tree
• Aggregate Report
• Graph Results
• View Results Table

View Results Tree

This report generates a request tree and clicks to view the information for each request.

Aggregate Report

This report generates aggregated statistics of requests, the main parameters being QPS, transmission speed, etc.

Graph Results

Powerful graphics report results

Some illustrations of graphical results

• No of Samples: Represents the number of requests sent to the server
• Deviation: Distribution of data representing the corresponding time changes of the server
• Latest Sample: Represents the time of the server's last request
• Throughtput: Here's how much data the server processes per minute
• Average: Represents the total run time divided by the number of requests sent to the server
• Median: Represents that half of the server time is below that value and the other half is above that value.

Some analysis based on graphical results

1. Throughput is relatively low at the start. As the number of requests increases, the throughput increases first and then decreases.
2. The deviation value is good. Keep a very stable state. If the number of deviations increases with the number of requests, the server becomes more and more unstable.