Diagnose Linux Performance in 60 Seconds: Essential Commands and Metrics Explained

Linux 60‑second analysis

The following checklist works for any performance problem and reflects the typical steps taken within the first minute after logging into a sluggish Linux host. uptime – shows current time, system run time, logged‑in users and the three load‑average values (1, 5 and 15 minutes). dmesg | tail or cat /var/log/messages – displays the last kernel messages; look for errors that may affect performance. vmstat 1 – reports processes, memory, paging, block I/O, traps, CPU activity. Important columns: r (runnable processes), b (blocked), free, si/so (swap in/out), us, sy, id, wa, st. mpstat -P ALL 1 – breaks CPU time per core into user, nice, system, iowait, irq, softirq, steal, guest and idle percentages. pidstat 1 – shows CPU usage per process (user and system time) and helps identify the most demanding processes. iostat -xz – provides I/O statistics for each block device. Key fields: r/s, w/s, rkB/s, wkB/s, await (average response time), avgqu‑sz (average queue length) and %util (device utilization). free -m – displays memory usage in megabytes; check the available column to see how much RAM is truly free. sar -n DEV 1 – network device statistics (packets/second, kilobytes/second, errors, drops). sar -n SOCK 1 – socket usage (total, TCP, UDP, RAW sockets). sar -n TCP,ETCP 1 – TCP metrics such as active/passive connections, segments in/out, retransmissions and resets. top – interactive view of processes, CPU and memory usage; press t or m for visual graphs.

Key metric interpretations

Load average < 0.7 × CPU count → light load; > CPU count → high load, possible bottleneck. iowait high → CPU spends time waiting for disk I/O; investigate storage performance.

High retrans/s or orsts/s → network issues or remote host problems. %util > 60 % on a disk often degrades performance; near 100 % indicates saturation.

Sample command output

uptime
 11:14:26 up 16 min, 1 user, load average: 21.37, 18.64, 16.25

vmstat 1
 r  b   swpd   free   buff  cache   si   so    bi    bo   in   cs us sy id wa st
 2  0      0 1057540   2072 1761012   0   0    68    27 1501  595  8  4 87  1  0

mpstat -P ALL 1
 all   12.89  0.00  14.84  50.00  6.25  4.30  0.00  0.00  0.00 11.72

iostat -xz
 avg-cpu:  %user   %nice %system %iowait  %steal %idle
          10.78    0.00   36.46   20.39    0.00   32.37
Device            r/s   w/s   rkB/s   wkB/s  await  %util
nvme0n1         101.86 23.92 4143.65  729.07   4.35   64.21

top
%Cpu(s): 0.9 us, 1.1 sy, 0.0 ni, 96.8 id, 0.0 wa, 0.6 hi, 0.5 si, 0.0 st

When iostat reports 100 % utilization, the device is busy for the entire sampling interval; it may still accept more load if the underlying hardware can handle parallel requests.

By following this 60‑second checklist you can quickly pinpoint whether a problem lies in CPU saturation, memory pressure, I/O bottlenecks or network issues, and then dive deeper with the appropriate tools.