Jmeter series (38)-Detailed performance monitoring tool nmon

If you want to learn Jmeter from scratch, you can check out this series of articles

https://www.cnblogs.com/poloyy/category/1746599.html

Preface##

nmon official introduction##

Data types that can be monitored by nmon##

Features of nmon##

  1. Nmon running ** will not occupy too much** system resources, under normal circumstances ** CPU utilization ** will not exceed 2%
  2. Powerful functions, comprehensive monitoring data types
  3. Combined with meter diagrams such as grafana, the monitored data can be displayed more intuitively in real time
  4. Good portability and compatibility
  5. Unattended monitoring
  6. You can view the monitoring results in real time, or you can view the monitoring results locally after the test is over

Disadvantages of nmon when doing performance testing##

Can not be integrated into Jmeter like ServerAgent to view monitoring results

Install and run nmon under Linux

Check the installation environment###

# View operating system information
uname -a 

# View linux release version
lsb_release -a

Download nmon

method one####

Official address: http://nmon.sourceforge.net/pmwiki.php?n=Site.Download

Upload to the server directory via FTP

Way two####

The server directly hits the command line

wget http://sourceforge.net/projects/nmon/files/nmon16e_mpginc.tar.gz

The latest version cannot be downloaded in this way

Baidu cloud download link

Contains nmon installation package and analysis data toolkit

Link: https://pan.baidu.com/s/1WHmlf-Em3FfKbUUa_GWi6A Extraction code: w6gy

Unzip###

tar -zxvf nmon16m_helpsystems.tar.gz

View the unzipped file###

Run nmon

My system is x86_64 centos8, knock

. /nmon_x86_64_centos8

Commonly used nmon shortcut commands###

Press multiple shortcut commands to display multiple monitoring results at the same time

nmon command line parameters##

When starting nmon, you can also specify some parameters

Parameters Function
- f The monitoring result is output as a file, default: machine name_date_time.nmon
- F Same function as -f, but specify the output file name
- s Sampling frequency, in seconds
- c Sampling times
- m nmon file save directory
- T Display processes with high resource usage

Command line chestnut####

. /nmon -f -t -s 10-c 6-m /data/nmon

Collect system resource data every 10s, 6 times in total (6 times in 1 min), and generate the final nmon file in the /data/nmon directory

Transfer this .nmon file to the local for subsequent result analysis

Focus

View nmon monitoring results##

Introduction to nmon_analyser###

Download nmon_analyser

Official: http://nmon.sourceforge.net/pmwiki.php?n=Site.Nmon-Analyser

Baidu cloud link: https://pan.baidu.com/s/1WHmlf-Em3FfKbUUa_GWi6A Extraction code: w6gy

After decompression, there are two files

Open nmon analyser v66.xlsm

Remember to open it in excel through office or wps, and need to start the macro

Resolve the failure of macro startup

Open .nmon file###

Click Analyze nmon data, select nmon file

View monitoring result data###

Graph analysis of nmon monitoring results##

The analysis result file is an excel table, which is divided into multiple tab pages such as SYS_SUMM, CPU_ALL, MEM, DISK_SUMM, etc. Each tab page has detailed data monitored under the topic and corresponding statistical charts

Each tab page displays detailed monitoring data for a certain indicator such as CPU, memory, network, and I/O

System summary interface: SYS_SUMM

Page display items

Parameters Description
Avg disk tps during an interval The average number of disk I/Os in the collection interval This value is equal to the average value of the IO/sec column in Sheet DISK_SUMM
Max disk tps during an interval The maximum number of disk I/Os in the collection interval This value is equal to the maximum value of the IO/sec column in Sheet DISK_SUMM
Max disk tps interval time Display the time point of the maximum disk I/O
Total number of Mbytes read The total number of megabytes read from the disk in the acquisition interval
Total number of Mbytes written Total number of Mbytes written to disk in the acquisition interval
Read/Write Ratio
User% Time percentage of all CPUs in User Mode in the acquisition interval (Avg, Max)
Sys% Time percentage of all CPUs in System Mode in the acquisition interval (Avg, Max)
Wait% The percentage of time that all CPUs are idle and waiting for I/O to complete in the collection interval (Avg, Max)
Idel% The percentage of time that all CPUs are idle during the acquisition interval (Avg, Max)
CPU% Total CPU usage (Avg, Max)

Focus

System Information: AAA

The information mainly includes: execute command, host CPU number, operating system kernel version information, host name and other information

System details: BBBP

Page information mainly includes: operating system version, host disk information, host CPU model, main frequency information, memory information, network card information, etc.

CPU usage: CPU_ALL, CPU_SUMM, CPU001...

Including host CPU usage summary and the running status of each CPU.

Focus on the CPU_ALL page here

On the CPU_ALL page, you can see the overall CPU data at each monitoring point as shown in the figure below, including User%, Sys%, Wait%, Idle%, Steal%, CPU%, and CPUs (number of cpu) and corresponding charts

CPU-ALL

Parameter Description####

Parameters Description
CPU Total Execution interval time list
User% Time percentage of all CPUs in User Mode in the acquisition interval (Avg, Max)
Sys% Time percentage of all CPUs in System Mode in the acquisition interval (Avg, Max)
Wait% The percentage of time that all CPUs are idle and waiting for I/O to complete in the collection interval (Avg, Max)
Idel% The percentage of the time that all CPUs are idle in the collection interval (Avg, Max). This value and the sum of User%, Sys%, Wait%, and Steal% are equal to 1
CPU% The overall CPU usage, this value is usually equal to User%+Sys%
CPUs Number of CPU cores, that is, how many cores does the operating system have

Disk read and write situation###

There are six tabs to view the disk status

Disk read, write and I/O statistics, read and write status of each disk partition of the system. Focus on DISK_SUMM here

DISK_SUMM

DISK_SUMM displays the Read/Write rate (KB/s) of all disks and partitions and the I/O rate of all disks and partitions according to the acquisition time.

Parameter Description####

Parameters Description
Disk total KB/s Execution interval time list
Disk Read KB/s The read rate of the disk device in the acquisition interval
Disk Write KB/s The write rate of the disk device in the acquisition interval
IO/sec The overall average disk IO per second in the collection interval

Memory situation: MEM

Parameter Description####

Parameters Description
memtotal Total size of physical memory, in MB
swaptotal The total size of virtual memory (ie swap space)
memfree Remaining physical memory size
swapfree Remaining virtual memory size
cached Cache size occupied by cache in physical memory
active Memory size in active use
buffers The size of file system buffers
swapcached The size of cache occupied by cache in virtual memory
inactive Infrequently used memory size

System network situation: NET, NETPACKET

Reflect the network operation of the system, the number of data packets read and written by each network adapter of the system

The NET page displays the data transfer rate of each network adapter in the system (kilobytes/second)

Parameter Description####

Parameters Description
Network I/O Execution interval time list
Total-Read The total size of data packets received by the network adapter per second (KB/s)
Total-Write(-ve) The total size of data packets sent by the network adapter per second (KB/s)

System process: PROC

Reflects the average number of running threads in the system and threads waiting to be switched

Collection time: ZZZZ

Record the time point when the nmon tool collects system information

Recommended Posts

Jmeter series (38)-Detailed performance monitoring tool nmon