Linux Performance Analysis for Beginners: A Quick Guide

If your Linux system is running slow, how do you find out what's wrong? This guide will help beginners analyze Linux performance step by step using simple tools. In just a minute, you can check CPU, memory, disk, and network performance using a few basic commands.

1. Check System Load (1 Second)

Run:

uptime

Output:

 10:45:23 up 5 days,  2:34,  2 users,  load average: 0.54, 0.66, 0.72

This shows how long your system has been running and the load average. The last three numbers represent the load over the last 1, 5, and 15 minutes. A high number (greater than the number of CPU cores) means your system is overloaded.

Troubleshooting Scenario:

Issue: Your system feels sluggish, and the load average is high. Solution: Check which processes are using the CPU.

2. Check for System Errors (1 Second)

Run:

dmesg | tail

Output:

[12345.678901] usb 1-1: USB disconnect, device number 3
[12345.679012] usb 1-1: new full-speed USB device number 4 using xhci_hcd

This shows the latest system messages, including hardware errors or crashes. If you see repeated errors, something might be wrong with your hardware or drivers.

Troubleshooting Scenario:

Issue: Your system crashes or behaves strangely. Solution: Check for error messages in dmesg and investigate further.

3. Check Overall System Performance (1 Second)

Run:

vmstat 1

Output:

procs -----------memory---------- ---swap-- -----io---- -system-- ------cpu-----
 r  b   swpd   free   buff  cache   si   so    bi    bo   in   cs us sy id wa st
 1  0      0  50000  10000 200000    0    0     5     2  200  500  5  1 94  0  0

This updates every second and shows:

CPU usage
Memory usage
Disk activity
System tasks

If the r (running tasks) value is high, your system is overloaded.

Troubleshooting Scenario:

Issue: System is slow, but you don't know if it's CPU, memory, or disk-related. Solution: Use vmstat to see which resource is overloaded.

4. Check CPU Usage Per Core (1 Second)

Run:

mpstat -P ALL 1

Output:

Linux 5.4.0-91-generic (ubuntu) 
10:45:23 CPU    %usr   %nice    %sys %iowait %irq   %soft  %steal  %guest   %idle
10:45:23 all    10.5    0.0     5.2    0.5   0.0    0.2    0.0      0.0    83.6

This shows CPU usage for each core every second.

If one core is much busier than the others, the workload is not balanced.
High %usr means applications are using a lot of CPU.
High %sys means the system itself is using too much CPU.

Troubleshooting Scenario:

Issue: One CPU core is always at 100%. Solution: Check which process is using it with top or pidstat.

5. Check CPU Usage by Process (1 Second)

Run:

pidstat 1

Output:

Linux 5.4.0-91-generic (ubuntu)   02/27/2025
10:45:23      UID       PID    %usr %system  %guest    %CPU   CPU  Command
10:45:23      1000      1234    10.0    2.5      0.0   12.5     1   firefox

This updates every second and shows how much CPU each process is using.

Troubleshooting Scenario:

Issue: The system is slow, and you need to find out which process is the cause. Solution: Run pidstat to find the most CPU-hungry process.

6. Check Disk Usage (1 Second)

Run:

iostat -xz 1

Output:

Device            r/s     w/s     rkB/s     wkB/s   await   svctm  %util
sda              1.24    2.45    100.23    200.56    5.23    1.02   7.5

This shows disk performance:

High %util means the disk is very busy.
High await means disk response time is slow.

Troubleshooting Scenario:

Issue: Applications are slow to load or save files. Solution: Check if the disk is overloaded with iostat.

7. Check Memory Usage (1 Second)

Run:

free -m

Output:

              total        used        free      shared  buff/cache   available
Mem:          16000        8000        5000        1000        3000       7000

This shows how much memory is used and available.

Troubleshooting Scenario:

Issue: The system is slow, and programs are freezing. Solution: Check if memory is running low with free -m.

8. Check Network Usage (1 Second)

Run:

sar -n DEV 1

Output:

10:45:23    IFACE rxpck/s txpck/s rxkB/s txkB/s
10:45:23     eth0   120.3   130.4   2000.5  1500.3

Troubleshooting Scenario:

Issue: Internet is slow or lagging. Solution: Use sar -n DEV to check if your network is overloaded.

9. Check Active Processes and CPU Usage (Real-Time)

Run:

top

Output:

top - 10:45:23 up 5 days,  2:34,  2 users,  load average: 0.54, 0.66, 0.72
Tasks: 250 total,   2 running, 248 sleeping,   0 stopped,   0 zombie
...
  PID USER      PR  NI    VIRT    RES    SHR S  %CPU %MEM     TIME+ COMMAND
 1234 user      20   0  200000  50000  10000 R  12.5  5.0   0:05.32 firefox

10. Capture Short-Lived High-CPU Processes

Run:

perf record -F 99 -a -g -- sleep 10

This command captures system performance data for 10 seconds at 99Hz, helping to identify processes that appear briefly but consume high CPU.

Troubleshooting Scenario:

Issue: A process appears in top but disappears too quickly to identify. Solution: Use perf record to capture the activity over time, then analyze the results using perf report.

Troubleshooting Scenario:

Issue: The system is slow, and you need to find the cause. Solution: Run top and look at the highest CPU-consuming processes.