free

What it does

free displays a snapshot of the system's memory usage: total, used, free, shared, buff/cache, and available RAM, plus swap usage, all sourced from /proc/meminfo. It's the fastest command-line way to check whether a Linux system is under memory pressure, though correctly reading its output requires understanding the difference between "free" and "available" memory.

Beginner examples

• free — show memory usage in kilobytes
• free -h — show memory usage in human-readable units (KiB/MiB/GiB)
• free -m — show memory usage in megabytes
• free -t — add a combined total line for memory + swap

free -h

Advanced examples

• Monitor memory usage live, refreshing every 2 seconds: free -h -s 2
• Limit a live monitor to 5 refreshes: free -h -s 2 -c 5
• Check swap usage specifically to see if the system has started swapping under pressure: look at the Swap: row's used column
• Script an alert when available memory drops below a threshold: free -m | awk '/^Mem:/{if ($7 < 500) print "low memory!"}'
• Compare available vs free to understand real memory headroom on a busy server.

free -m | awk '/^Mem:/{print "Available: " $7 " MB"}'

Common mistakes

• Panicking over a low number in the free column while ignoring available — Linux intentionally uses spare RAM for disk cache (shown in buff/cache), which is reclaimed instantly when applications need it.
• Interpreting high swap used values in isolation — some swap usage is normal and can persist even after memory pressure eases, since the kernel doesn't always swap pages back in until they're touched again.
• Forgetting that older versions of free (pre-procps 3.3.10) don't have an available column at all, making used/free harder to interpret correctly on very old systems.
• Comparing raw byte counts across systems without normalizing units — always use -h, -m, or -g for consistent, readable comparisons.

Tips

• Trust the available column, not free, when deciding whether a system is actually low on memory.
• Use free -h -s 2 as a lightweight live memory monitor when htop/top isn't installed or is overkill for a quick check.
• Pair free with vmstat 1 when diagnosing whether high memory usage correlates with active swapping (look at si/so columns in vmstat).

Best practices

• Build memory alerting around the available column, not free, to avoid false alarms caused by normal disk cache usage.
• When investigating a suspected memory leak, track used and available over time (e.g. with free -s) rather than relying on a single snapshot.
• Always specify a unit flag (-h, -m, or -g) in documentation or scripts so output is unambiguous across different free versions and locales.

Real-world use cases

• Quickly diagnosing whether an application slowdown is caused by memory exhaustion.
• Checking whether a server has started swapping heavily, which can severely degrade performance.
• Deciding whether a VM or container needs a memory limit increase based on sustained available memory being low.
• Including a memory snapshot in incident reports or capacity-planning documentation.

Common interview questions

• What's the difference between "free" and "available" memory in free's output? "Free" is completely unused memory; "available" is an estimate of memory that can be given to new applications without swapping, including reclaimable cache/buffers — "available" is the more meaningful number in practice.
• Why does a healthy Linux server often show very little in the free column? Because Linux uses spare RAM for disk cache/buffers to speed up I/O, and that memory is instantly reclaimable — it's not wasted or leaked.
• How would you monitor memory usage continuously from the command line? free -h -s <seconds> refreshes the report at the given interval; -c can cap the number of refreshes.