How to Pinpoint CPU‑Hogging Processes in Under 2 Minutes

CPU Full‑Utilization Troubleshooting

In production, a sudden CPU spike can degrade service. The following method identifies the responsible Java process and thread within about two minutes.

Step‑by‑step workflow

Identify the CPU‑hungry process (≈30 s) Run top , press P to sort by CPU usage, and note the PID of the process with the highest %CPU.

Find the offending thread inside that process (≈30 s) Execute top -Hp <PID> (‑H shows threads, ‑p selects the process). Record the TID of the top‑ranked thread (displayed in decimal).

Convert the thread ID to hexadecimal (≈10 s) Java stack traces use hexadecimal thread IDs (nid). Convert the decimal TID with printf "%x\n" <TID> . The output is the hex ID needed for the next step.

Locate the code in the Java stack trace (≈50 s) Run jstack <PID> | grep <hexTID> -A 30 and examine the surrounding lines.

If the thread state is RUNNABLE, the displayed code line is actively executing (e.g., infinite loop or heavy computation).

If the thread name is VM Thread or GC task thread, frequent Full GC may be the cause of the CPU surge.

Following this disciplined approach lets you pinpoint the exact process or Java thread responsible for a CPU overload quickly, enabling rapid remediation.