How to check memory usage in Linux

Monitoring memory usage on Linux keeps applications responsive and reduces the chance of sudden out-of-memory conditions that can terminate important services. Understanding how much RAM and swap space are in use makes it easier to troubleshoot slow systems, size workloads correctly, and spot memory leaks before they become critical.

The kernel exposes memory statistics through the /proc/meminfo interface and makes them available to tools such as free, top, htop, and vmstat. These utilities summarize total and available memory, show how much is used for page cache and buffers, and reveal whether the system is relying on swap space. Most of them are available by default on common Linux distributions.

Graphical utilities like GNOME System Monitor or KDE System Monitor provide similar information in a desktop-friendly way, but terminal-based tools remain the standard choice on servers and remote systems where no GUI is present. Reading memory statistics is safe, but some commands benefit from sudo access to see all processes, and high memory pressure visible in these outputs should be treated as a sign to optimize workloads or upgrade resources.

1. Open a terminal on the Linux system with an account that can run administrative commands when needed.
2. Display overall memory usage using the free command.

   $ free
                 total        used        free      shared  buff/cache   available
   Mem:        2013428      999352      113610        6188      900376      842184
   Swap:        969960         524      969436

   This output summarizes physical RAM and swap, including memory used for page cache and buffers.

3. View memory usage in a human-readable format.

   $ free -h
                 total        used        free      shared  buff/cache   available
   Mem:          1.9Gi       976Mi       111Mi       6.0Mi       879Mi       822Mi
   Swap:         947Mi       0.0Ki       946Mi

   The -h option prints sizes in GiB and MiB, which is easier to interpret than raw kilobytes.

4. Monitor real-time memory usage using the top command.

   $ top
   top - 16:25:49 up 40 min,  2 users,  load average: 0.08, 0.23, 0.34
   Tasks: 281 total,   1 running, 280 sleeping,   0 stopped,   0 zombie
   %Cpu(s):  5.9 us, 17.6 sy,  0.0 ni, 76.5 id,  0.0 wa,  0.0 hi,  0.0 si,  0.0 st
   MiB Mem :   1966.2 total,    108.4 free,    975.9 used,    881.9 buff/cache
   MiB Swap:    946.9 total,    0.5 used,    946.4 free.   842.1 avail Mem

   Press q to exit top; the MiB Mem line shows total, free, used, and cached memory in real time.

5. Run htop for an interactive, colorized view of per-process memory usage.

   $ htop

   htop presents CPU and memory meters at the top and allows sorting by columns such as RES and VIRT using function keys.

6. Check detailed memory statistics from the kernel using the /proc file system.

   $ cat /proc/meminfo
   MemTotal:       15712000 kB
   MemFree:        10035200 kB
   MemAvailable:   12567000 kB
   Buffers:          210000 kB
   Cached:          1800000 kB
   ##### snipped #####

   Fields such as MemTotal, MemAvailable, Buffers, and Cached explain how physical memory is divided between applications and cache.

7. View memory and paging activity using the vmstat command.

   $ vmstat -s
        15712000 K total memory
         2350500 K used memory
         9847200 K active memory
   ##### snipped #####

   vmstat adds information about swap and paging, which helps detect thrashing and sustained swap activity.

8. Sort running processes by memory usage using the ps command.

   $ ps aux --sort=-%mem | head
   USER       PID %CPU %MEM    VSZ   RSS TTY      STAT START   TIME COMMAND
   root      1234  5.1 10.5 123456 23456 ?       Ssl  10:00   2:30 process1
   ##### snipped #####

   Sorting by %MEM surfaces the heaviest consumers of RAM so that tuning can focus on the most demanding processes.

9. Compare MemTotal from /proc/meminfo with totals reported by free or top to confirm that memory statistics are consistent.

   $ grep MemTotal /proc/meminfo
   MemTotal:       2013428 kB

   A small difference between tools is normal due to rounding and units, but large discrepancies can indicate measurement misunderstandings such as including or excluding cache.