Skip to content

top

Using top will show the current status of the system.
Its output is very detailed.

top Output

First Line

In the first line we see something similar to this: 

top - 19:38:28 up 2 days, 20:47, 0 users, load average: 0.52, 0.58, 0.59

  • top: program name
  • 19:38:28: Current system time.
  • up 2 days, 20:47: System Uptime. The time from the last boot.
  • 0 users: number of active users. Here we can see similar information to the command who.
  • The last part is the load average. It's very important, and very often misunderstood. 
    load average: 0.52, 0.58, 0.59
    These numbers show the load average for the system in the last 1 min, 5 mins, and 15 minutes respectively.
    • The load average shows the average number of processes both running and waiting for CPU time.
    • The number of CPUs (as well as their cores && threads) should be considered when evaluating the load averages.
      • For example, a load average of 10 with a 1-core CPU means that your system is overloaded, but the same load average with a 12-core CPU is not overloaded.

Second Line

The second line shows information about processes in the system. 

Tasks: 6 total, 1 running, 5 sleeping, 0 stopped, 0 zombie

  • total: Shows all processes in the system
  • running: Currently active processes.
    • This means these processes are using CPU right now
  • sleeping: Generally - process is waiting for something.
    • It may be I/O operation for example.
  • stopped: Stopped processes (e.g., with <C-z>)
  • zombie: Very important state to understand.
    • A zombie is a process that has finished its job but still exists in the process table.
    • In simple terms, zombie processes are waiting for an exit signal.
    • This can happen when the parent process fails for whatever reason.
    • Sometimes you can kill a zombie process by killing the parent process, but this is usually not the case.

Third Line

This line shows the CPU(s) utilization, splitted to specific types. Let's go through them one by one. 

%Cpu(s): 13.9 us, 9.5 sy, 0.0 ni, 76.3 id, 0.0 wa, 0.4 hi, 0.0 si, 0.0 st

  • us: user - All user processes are combined in this number.
    • So, our sessions too.
  • sy: system - processes owned by system (kernel)
  • ni: nice - this is important to understand.
    • nice allows us to change the priority of the process.
    • The standard value for processes is 0 , but we can modify it from 19 (lowest) to -20 (highest) priority.
    • This statistic here shows all processes with the niceness set abow 0. So, the processes which will be executed by the system, when "systemm will have time for it".
  • id: idle - idle time means that the system is bored and do nothing.
  • wa: iowait - the number repspresents the time (which is a subset of idle time) when the process is waiting for input/output operation.
    • This statistic is very important, because it may show the issue outside the CPU, in other hardware (but not only) components.
  • hi: hardware interrupts.
    • These are physical interrrupts from hardware and are handled by CPU itself.
  • si: software interrupts. These are generated by software and are handled by kernel.
  • st: steal time - very important to understand, especially when we are working on virtualized environment.
    • This number represents the time "stealed" from the virtual machine by hypervisor.
    • Another words, how long our system needs to wait for resources from hypervisor.

Fourth and Fifth Lines

Both of these lines represent the memory information.
The difference is that the first line is physical memory and the second is swap.

MiB Mem :  16217.5 total,   6184.9 free,   9808.7 used,    224.0 buff/cache
MiB Swap:  49152.0 total,  48436.2 free,    715.8 used.   6278.3 avail Mem

  • total, free, used: obvious.
  • buff/cache: A combined value of buffer memory used by the kernel and cache, as well as buffer memory used by the page cache.
  • available: means that any newly started program can use, at maximum, this amount of memory.

Processes list

Below the first five lines is the process list.

This list contains the following fields:

  • PID: Process ID number.
    • It is a unique number given to the process by the system.
  • USER: Process owner.
    • The process was started by this user.
  • PR: Default priority of the process, scheduled by the kernel when the process was started.
  • NI: Nice.
    • If nice was performed against the process.
  • VIRT: Total amount of memory used by the process.
  • RES: RAM memory used by process.
  • SHR: Amount of memory shared with other processes.
  • S: Process state.
  • %CPU: Represents CPU usage.
  • %MEM: Represents memory usage.
  • TIME+: Total time of CPU usage by the process.
  • COMMAND: The command that started the process.