Master JVM Performance: Essential Tools (jps, jstack, jmap, jstat, hprof)

Introduction

DK provides many convenient JVM performance tuning and monitoring tools. Besides integrated tools like VisualVM and jConsole, there are lightweight utilities such as jps, jstack, jmap, jhat, jstat, and hprof. This article introduces these tools to help developers understand and address common JVM performance problems.

Typical Enterprise Java Issues

OutOfMemoryError – insufficient memory

Memory leaks

Thread deadlocks

Lock contention

Excessive CPU usage by Java processes

These problems are often overlooked; developers may simply restart the server or increase memory without investigating the root cause. Understanding and solving them is essential for Java engineers.

A. jps (Java Virtual Machine Process Status Tool)

jps displays the status of JVM processes. Basic syntax: jps [options] [hostid] Options:

-q : suppress class name, JAR name, and main arguments

-m : show main arguments

-l : display the full package name of the main class or JAR

-v : display JVM arguments

Example output:

root@ubuntu:/# jps -m -l
2458 org.artifactory.standalone.main.Main /usr/local/artifactory-2.2.5/etc/jetty.xml
29920 com.sun.tools.hat.Main -port 9998 /tmp/dump.dat
3149 org.apache.catalina.startup.Bootstrap start
...

B. jstack

jstack shows thread stack traces of a Java process. Syntax: jstack [option] pid Key options:

-l : long listings, prints additional lock information (useful for deadlock analysis)

-m : mixed mode, includes native stack frames

Example workflow to find the most CPU‑intensive thread:

Identify the Java process ID (e.g., ps -ef | grep mrf-center).

Find the thread with highest CPU time using top -Hp <pid>.

Convert the thread ID to hexadecimal and grep it in the jstack output.

Sample commands and output:

root@ubuntu:/# ps -ef | grep mrf-center | grep -v grep
root 21711 1 1 14:47 pts/3 00:02:10 java -jar mrf-center.jar

root@ubuntu:/# printf "%x
" 21742
54ee

root@ubuntu:/# jstack 21711 | grep 54ee"PollIntervalRetrySchedulerThread" prio=10 tid=0x00007f950043e000 nid=0x54ee
in Object.wait() [0x00007f94c6eda000]

The CPU consumption was in PollIntervalRetrySchedulerThread waiting on Object.wait(). The corresponding source code snippet is shown below:

C. jmap and jhat

jmap inspects heap memory usage, often used together with jhat for analysis. jmap [option] pid Common options:

-permstat pid : prints class loader and permanent generation statistics

-heap pid : shows heap configuration and usage

-histo[:live] pid : prints a histogram of object counts and sizes (live only if :live is added)

-dump:format=b,file=<file> pid : dumps the heap to a binary file

Example dump command:

root@ubuntu:/# jmap -dump:format=b,file=/tmp/dump.dat 21711
Dumping heap to /tmp/dump.dat ...
Heap dump file created

The dump can be examined with tools like MAT, VisualVM, or jhat:

jhat -J-Xmx512m -port 9998 /tmp/dump.dat

D. jstat

jstat provides JVM statistics.

jstat [generalOption|outputOptions] vmid [interval[s|ms] [count]]

Example showing GC statistics every 250 ms for four samples:

jstat -gc 21711 250 4

Key columns explained (S0C, S1C, EC, OC, PC, YGC, YGT, FGC, FGCT, GCT, etc.) after understanding the JVM heap layout (young generation = Eden + two Survivor spaces; old generation; permanent generation).

E. hprof (Heap/CPU Profiling Tool)

hprof can profile CPU usage and heap memory.

java -agentlib:hprof=cpu=samples,interval=20,depth=3 Hello
java -agentlib:hprof=cpu=times Hello.java
java -agentlib:hprof=heap=sites Hello.java
java -agentlib:hprof=heap=dump Hello.java

Sampling mode records CPU usage every 20 ms with a stack depth of 3, generating java.hprof.txt. Times mode provides finer‑grained CPU data via bytecode injection. Heap site mode records allocation sites, while heap dump mode produces a detailed heap dump. Using -Xrunprof:heap=sites in production is discouraged due to high overhead.