jmeter monitoring Linux servers

 

 

 

　　Recent performance tests, loadrunner compare pits, all kinds of compatibility issues, no jmeter really easy to use, the project done, where records about the use of jmeter during pit of it, to do the project after the accumulation of experience points.

　　jmeter how to configure the monitoring linux Needless to say, quite simply, a lot of online tutorials direct cut to the chase.

　　The project implementation process has two interfaces, one for landing, upload a file, this is the background.

problem:

　　1, you need to log in to upload test interfaces, this time can only modify the script, the login request into a separate thread group, upload request a separate thread group, set up the session after landing as global variables can upload thread group use.

　　2, type a file upload request packet capture tool can be used to crawl, in addition to File Upload set up, but also the parameters set.

　　3, server monitoring

　　　Most jmeter server monitoring online are the same, do not have the various parameters indicators speak very clearly, explain in detail below

　　　

　　Memory is a percentage

　　That shows the percentage of CPU

　　Hard disk read and write times are displayed, but only to read and write under the monitoring / home directory (this is really too pit, and there is no data in the hard disk when doing the project monitoring, server monitoring can only be carried out using the nmon)

　　Eth0 network is displayed

 But that can be set:

1. CPU：
   • For each index entry values are representative of a percentage , such as the default configuration ( combined) seen in the graph at a certain time value is 30, it means that at this time the total cpu time accounted for 30%.
   • There are two more useful place worthy of note: First Scope area by Per Process option can be acquired CPU usage of the specified process , and second in CPU Cores area, we can select individual Core surveillance specified.
2. Memory：
   • Each index entry, usedperc(default) and freepercthe value of two represents the percentage of the total memory , the remaining value of index entries are referred to as memory size , select the corresponding think, you can see Metric Unitthe region configuration unit becomes available, it would be more usually Mb for observation.
   • Also, you can select the data to monitor the specified process
3. I Disk / O :

• • Each index entry, queue(default) value represents waiting for I / O queue length reads, writesrepresenting per second read / write cycles readbytes, writebytesas the name implies, indicates the amount of data per second read / write unit in the same Metric Unitconfiguration area, usually Mb would be more suitable for observation.

　　4. Network

• • Monitoring the interface eth0 to obtain the second data packet transmission rate