Use top, htop, or ps aux --sort=-%cpu | head -10 to identify high-CPU processes by PID, command, and user. 2. Check system load with uptime and compare load average to CPU core count using nproc to determine if CPU contention exists. 3. Analyze CPU usage types with vmstat 1 5 (us, sy, id, wa) and use mpstat -P ALL 1 for per-core breakdown to detect saturation or I/O waits. 4. Investigate the root cause by examining the process command line with ps, checking threads via top -H -p , tracing system calls with strace -p cautiously, and profiling with perf top -p for code-level insights. 5. Address common issues: restart apps with infinite loops, fix misconfigured cron jobs using crontab -l, scan for malware with chkrootkit, reschedule resource-heavy scripts, tune JVM settings for Java GC problems, and update outdated software. 6. Prevent recurrence by setting up monitoring with tools like netdata or Prometheus, configuring alerts, and optionally limiting CPU with cpulimit -p -l 50, ensuring sustained high usage is either expected or resolved through systematic observation, identification, analysis, and resolution.

High CPU usage in a Linux environment can slow down system performance, affect application responsiveness, and even lead to service outages. Identifying and resolving the root cause requires a methodical approach using built-in tools and system monitoring techniques. Here’s how to troubleshoot it effectively.

1. Identify the Processes Using the Most CPU

Start by pinpointing which processes are consuming excessive CPU resources.

Use top or htop (if installed):

top

In the top interface:

• Press P to sort processes by CPU usage.
• Look at the %CPU column to identify the top consumers.
• Note the PID, user, and command name.

Alternatively, use htop for a more user-friendly, color-coded view:

htop

For a one-line summary of top CPU-consuming processes:

ps aux --sort=-%cpu | head -10

This shows the top 10 processes by CPU usage.

2. Check System-Wide CPU Utilization and Load Average

Use uptime or top to see the system load average:

uptime

Output example:

14:20:03 up 10 days,  2:15,  1 user,  load average: 2.45, 1.80, 1.20

• Load average values represent the number of processes waiting for CPU time over 1, 5, and 15 minutes.
• If the 1-minute load exceeds the number of CPU cores, you may have a CPU bottleneck.

Check the number of CPU cores:

nproc

Compare load average to core count. For example, on a 4-core system, a load average of 6 indicates contention.

3. Analyze CPU Usage by Type with vmstat or mpstat

Use vmstat to get a quick overview of CPU usage breakdown:

vmstat 1 5

This samples every second, 5 times. Look at the last few columns:

• us (user): CPU used by user processes.
• sy (system): CPU used by kernel/system processes.
• id (idle): Idle time.
• wa (iowait): CPU waiting for I/O.

High sy may indicate kernel or system call overhead.
High wa suggests I/O bottlenecks (though not CPU directly, it can cause processes to wait and re-schedule).

For per-CPU analysis, use mpstat (from the sysstat package):

mpstat -P ALL 1

This shows detailed CPU usage across all cores and helps detect imbalances or single-core saturation.

4. Investigate the Root Cause of High CPU Usage

Once you’ve identified a high-CPU process, dig deeper:

a. Check the Process Command Line

ps -p <PID> -o pid,ppid,cmd,%cpu,%mem

This shows what the process is actually doing.

b. Inspect Process Threads

Some applications use multiple threads. Use:

top -H -p <PID>

This shows per-thread CPU usage. A single rogue thread can spike CPU.

c. Trace System Calls with strace (Use Cautiously)

strace -p <PID>

This shows system calls the process is making. Useful to see if it’s stuck in a loop or making excessive calls.

?? Warning: strace adds overhead—avoid in production unless necessary.

d. Profile with perf (Advanced)

perf top -p <PID>

Shows real-time function-level CPU usage. Great for identifying hot code paths in applications.

5. Common Causes and Fixes

Here are frequent culprits and how to address them:

• Infinite loops in applications: Restart or patch the app. Check logs for errors.
• Misconfigured cron jobs: Run crontab -l and check /etc/crontab for scripts running too frequently.
• Malware or unauthorized processes: Use chkrootkit or rkhunter to scan for rootkits.
• Resource-hungry scripts (e.g., backups, reports): Schedule them during off-peak hours.
• Java applications with garbage collection issues: Monitor GC logs; tune JVM settings.
• Outdated or buggy software: Update packages and apply patches.

6. Monitor and Prevent Recurrence

Set up monitoring to catch issues early:

• Use tools like netdata, prometheus node_exporter, or glances.
• Configure alerts for sustained high CPU usage.
• Regularly review system performance as part of maintenance.

You can also limit process CPU usage with cpulimit:

cpulimit -p <PID> -l 50

This limits the process to 50% CPU.

Basically, troubleshooting high CPU usage comes down to: observe → identify → analyze → resolve. Use the right tools in order, and always verify whether the usage is expected (e.g., batch processing) or truly abnormal.

The above is the detailed content of Troubleshooting High CPU Usage in a Linux Environment. For more information, please follow other related articles on the PHP Chinese website!