Information is essential to resolving any pc problem, including issues with or associated with Linux and the equipment on which it works. There are various tools designed for and incorporated with most distributions despite the fact that they aren’t all set up by default. These tools may be used to get large sums of information.

This article discusses a few of the interactive command line interface (CLI) tools that are given with or which may be easily installed on Red Hat related distributions including Red Hat Enterprise Linux, Fedora, CentOS, and other derivative distributions. Although there are GUI tools obtainable plus they offer good information, the CLI tools provide all the same information plus they are often usable because many servers don’t have a GUI interface but all Linux systems possess a command line interface.

This article specializes in the tools that I take advantage of typically. If I didn’t cover your favorite device, make sure you forgive me and why don’t we all know very well what tools you make use of and just why in the responses section.

My head to tools for problem determination in a Linux environment are nearly always the operational program monitoring tools. For me, they are top, atop, and htop.

All of these equipment monitor CPU and memory usage, & most of them list information regarding running processes at the very least. Some monitor other areas of a Linux system as well. All offer near real-time views of program activity.

Load averages

Before I continue to discuss the monitoring tools, it is necessary to go over load averages in greater detail.

Load averages are a significant requirements for measuring CPU utilization, but what will this really mean when We say that the 1 (or 5 or 10) minute load typical is 4.04, for example? Load typical can be viewed as a way of measuring demand for the CPU; it really is a true number that represents the common number of instructions looking forward to CPU time. Therefore this is a genuine measure of CPU overall performance, unlike the typical “CPU percentage” which include I/O wait times where the CPU is not actually working.

For example, a completely utilized single processor system CPU could have a load average of 1. Which means that the CPU is maintaining precisely with the demand; quite simply it has perfect utilization. Lots average of significantly less than one means that the CPU is definitely underutilized and a load average in excess of 1 implies that the CPU can be overutilized and that there surely is pent-up, unsatisfied demand. For instance, a load ordinary of just one 1.5 in one CPU program indicates that one-third of the CPU instructions are forced to wait to be executed before one preceding it has finished.

This is true for multiple processors also. If a 4 CPU system includes a load standard of 4 after that it has ideal utilization. If a load is had by it average of 3.24, for instance, then three of its processors are fully utilized and one is utilized in about 76%. In the example above, a 4 CPU system has a 1 minute load average of 4.04 and therefore there is absolutely no remaining capability among the 4 CPUs and some guidelines are forced to hold back. A perfectly utilized 4 CPU system would show lots average of 4.00 so that the program in the example is loaded but not overloaded fully.

The optimum condition for load average is for this to equal the full total number of CPUs in something. That would imply that every CPU is utilized yet no instruction should be forced to wait fully. The longer-term load averages offer indication of the entire utilization trend.


All of the monitors discussed here allow you to send signals to running processes. Each of these signals has a specific function though some of them can be defined by the receiving program using signal handlers.
The separate kill command can also be used to send signals to processes outside of the monitors. The kill -l can be used to list all possible indicators that can be sent. Three of these signals may be used to kill a process.
• SIGTERM (15): Signal 15, SIGTERM is the default signal sent by top and the other monitors when the k key is pressed. It may also be the least effective because the system must have a signal handler built into it. The program’s signal handler must intercept incoming signals and act accordingly. So for scripts, most of which do not have transmission handlers, SIGTERM is ignored. The idea behind SIGTERM is that by simply telling the program that you want it to terminate itself, it will take advantage of that and clean up things like open files and then terminate itself in a controlled and nice manner.
• SIGKILL (9): Signal 9, SIGKILL provides a means of killing even the most recalcitrant programs, including scripts and additional programs that have no signal handlers. For scripts and various other programs with no signal handler, however, it not only kills the running script but it also kills the shell session in which the script is operating; this may not be the behavior that you would like. If you want to kill a process and you don’t care about being good, this is the transmission you want. This signal cannot be intercepted by a sign handler in the program code.
• SIGINT (2): Signal 2, SIGINT can be used when SIGTERM does not work and you want this program to die a little more nicely, for example, without killing the shell session in which it is running. SIGINT sends an interrupt to the program in which the program is running. This is equivalent to terminating a running plan, particularly a script, with the Ctrl-C key combination.
To experiment with this, open a terminal session and create a file in /tmp named cpuHog and make it executable with the permissions rwxr_xr_x. Add the following content to the file.

#!/bin/bash # This little program is a cpu hog X=0;while [ 1 ];do echo $X;X=$((X+1));done

Open another terminal program in a different windows, position them next to each other so that you can watch the total results and run top in the brand new session. Run the cpuHog system with the next command:


This scheduled program simply counts up by one and prints the existing value of X to STDOUT. And it sucks up CPU cycles. The terminal program where cpuHog is operating should show an extremely high CPU use in top. Take notice of the effect it has on system overall performance in top. CPU usage should immediately go method up and the strain averages should also begin to increase over period. If you would like, you can open extra terminal sessions and begin the cpuHog plan in them so you have multiple situations running.

Determine the PID of the cpuHog program you want to destroy. Press the k essential and appearance at the message beneath the Swap line in the bottom of the overview section. Top asks for the PID of the procedure you would like to kill. Enter that press and PID Enter. Now top requests the signal shows and quantity the default of 15. Try each one of the signals described here and take notice of the total results.

3 open source tools for Linux system monitoring


Among the first tools I take advantage of when performing problem dedication is top. I love it because it ‘s been around since permanently and is always obtainable as the other tools might not be installed.

The top program is an extremely powerful utility that delivers a lot of information regarding your running system. This consists of data about memory utilization, CPU loads, and a listing of running processes like the amount of CPU memory and time being employed by each process. Top displays system info in near real-time, updating (by default) every three seconds. Fractional mere seconds are allowed by best, although very small values can place a substantial load the operational system. Additionally it is interactive and the info columns to be shown and the type column could be modified.

An example output from the very best system is shown in Figure 1 below. The output from top is split into two sections which are known as the “overview” section, which is the top portion of the result, and the “procedure” section which may be the lower part of the output; I shall use this terminology for top level, atop and htop in the curiosity of consistency.

The top program includes a number of useful interactive commands you may use to control the display of data also to manipulate individual processes. Use the h command to see a brief help web page for the many interactive commands. Make sure to press h to see both pages of the help twice. Utilize the q command to give up.

Summary section

The summary portion of the output from top is an overview of the operational program status. The first series shows the system uptime and the 1, 5, and 15 minute load averages. In the example below, the strain averages are 4.04, 4.17, and 4.06 respectively.

The next line shows the amount of processes active and the status of every currently.

The relative lines containing CPU figures are shown next. There can be a single line which combines the figures for all CPUs within the operational system, as in the example below, or one collection for each CPU; in the full case of the computer utilized for the example, this is an individual quad primary CPU. Press the 1 essential to toggle between your consolidated display of CPU use and the screen of the average person CPUs. The info in these relative lines is displayed as percentages of the full total CPU time available.

These and the various other areas for CPU data are described beneath.

  • us: userspace – Applications and other programs running in user space, i.e., not in the kernel.
  • sy: system calls – Kernel level functions. This does not include CPU time taken by the kernel itself, just the kernel system calls.
  • ni: nice – Processes that are running at a positive nice level.
  • id: idle – Idle time, i.e., time not used by any running process.
  • wa: wait – CPU cycles that are spent waiting for I/O to occur. This is wasted CPU time.
  • hi: hardware interrupts – CPU cycles that are spent dealing with hardware interrupts.
  • si: software interrupts – CPU cycles spent dealing with software-created interrupts such as system calls.
  • st: steal time – The percentage of CPU cycles that a virtual CPU waits for a real CPU while the hypervisor is servicing another virtual processor.

The last two lines in the summary section are storage usage. They show the physical memory usage including both swap and RAM space.

Process section

The process portion of the output from top is all of the running processes in the system-at least for the amount of processes that there is room on the terminal screen. The default columns shown by best are described below. Other columns can be found and each can be added with an individual keystroke usually. Refer to the very best man page for information.

  • PID – The Process ID.
  • USER – The username of the process owner.
  • PR – The priority of the process.
  • NI – The nice number of the process.
  • VIRT – The total amount of virtual memory allocated to the process.
  • RES – Resident size (in kb unless otherwise noted) of non-swapped physical memory consumed by a process.
  • SHR – The amount of shared memory in kb used by the process.
  • S – The status of the process. This can be R for running, S for sleeping, and Z for zombie. Less frequently seen statuses can be T for traced or stopped, and D for uninterruptable sleep.
  • %CPU – The percentage of CPU cycles, or time used by this process during the last measured time period.
  • %MEM – The percentage of physical system memory used by the process.
  • TIME+ – Total CPU time to 100ths of a second consumed by the process since the process was started.
  • COMMAND – This is the command that was used to launch the process.


I like atop also. It is a fantastic monitor to use if you want more details about this kind of I/O activity. The default refresh interval is usually 10 seconds, but this is often transformed using the interval i control to whatever is suitable for what you want to perform. atop cannot refresh at sub-second intervals like top can.

Utilize the h command to show help. Make sure to observe that there are multiple webpages of help and you may utilize the space bar to scroll right down to see the rest.

One good feature of atop is that it could save raw overall performance data to a document and play it back later on for close inspection. That is handy for searching for internmittent problems, especially ones that occur during occasions when you cannot monitor the system directly. The atopsar program can be used to play back again the info in the saved document.

Summary section

atop contains a lot of the same info as best but also shows information regarding network, raw disk, and logical volume activity. Remember that for those who have the horizontal display real-estate to aid a wider display, extra columns will be displayed. Conversely, when you have much less horizontal width, fewer columns are shown. I also like this atop displays the existing CPU frequency and scaling factor-something I’ve not really seen on any additional of these monitors-on the next collection in the rightmost two columns.

Process section

The atop process display includes a few of the same columns mainly because that for top, but it addittionally includes disk I/O information and thread count for each process and also virtual and true memory growth statistics for every process. Much like the summary section, extra columns will display when there is adequate horizontal display real-estate.

atop may also provide detailed information regarding disk, memory space, network, and scheduling info for each process. Simply press the d, m, n or s keys respectively to see that data. The g important returns the screen to the generic procedure display.

Sorting could be accomplished easily through the use of C to sort simply by CPU use, M for memory usage, D for disk utilization, N for network use and A intended for automatic sorting. Automatic sorting generally sorts procedures by the most occupied resource. The network usage can only become sorted if the netatop kernel module is usually installed and loaded.

You may use the k key to kill a process but there is absolutely no choice to renice an activity.

By default, network and disk products that no activity occurs throughout a given period interval aren’t displayed. This can result in mistaken assumptions about the equipment configuration of the sponsor. The f command may be used to pressure atop to show the idle resources.


The atop man page identifies global and user level configuration files, but none are available in my very own Fedora or CentOS installations. Addititionally there is no control to save lots of an altered configuration and a save will not happen automatically when this program is terminated. Therefore, there is apparently now way to create configuration changes permanent.


The htop program is similar to top but on steroids. It can look nearly the same as top, but it additionally provides some features that top will not. Unlike atop, nevertheless, it does not offer any disk, network, or I/O info of any type.

Summary section

The summary portion of htop is displayed in two columns. It is extremely flexible and may be configured with a number of different types of info in just about any order you prefer. Although the CPU utilization sections of best and atop can be toggled between a mixed display and a screen that presents one bar graph for each CPU, htop cannot. So that it includes a number of different alternatives for the CPU screen, including a single mixed bar, a bar for every CPU, and different combinations where specific CPUs could be grouped together right into a single bar.

I think that is a cleaner overview display than a few of the other program monitors in fact it is easier to go through. The drawback to the summary section is usually that some information isn’t available in htop that’s available in the additional monitors, such as for example CPU percentages by consumer, idle, and system period.

The F2 (Set up) key can be used to configure the overview section of htop. A listing of available data shows is demonstrated and you can make use of function keys to include them left or correct column also to move them along within the chosen column.

Process section

The process portion of htop is quite similar compared to that of top. Much like the other monitors, processes can be sorted some of several elements, including CPU or memory space usage, consumer, or PID. Remember that sorting is not feasible when the tree look at is selected.

The F6 key enables you to choose the sort column; it shows a listing of the columns designed for sorting and you decide on the column you need and press the Enter essential.

You may use the up and down arrow keys to select a process. To kill an activity, use the along arrow keys to choose the target procedure and press the k important. A list of indicators to send the procedure is displayed with 15, SIGTERM, selected. You can specify the transmission to use, if not the same as SIGTERM. You might utilize the F7 and F8 keys to renice the chosen process.

One command We especially like is F5 which displays the working procedures in a tree format rendering it easy to look for the parent/child associations of running processes.


Each user has their own configuration file, ~/.config/htop/htoprc and changes to the htop configuration are stored there automatically. There is no global configuration file for htop.


Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: