Comprehensive Guide to Troubleshooting CPU, Disk, Memory, GC, and Network Issues in Java Applications

Online service failures often involve CPU, disk, memory, and network problems; most incidents span multiple layers, so a systematic four‑step inspection—CPU → Disk → Memory → Network—is recommended, using tools like jstack, jmap, jstat, top, vmstat, iostat, netstat, and ss.

CPU : Identify high‑CPU threads with ps to get the PID, then top -H -p <pid> to find hot threads, convert the thread ID to hexadecimal ( printf '%x\n' <pid>), and locate the stack in a jstack dump using jstack <pid> | grep '<nid>' -C5 --color. Focus on WAITING/TIMED_WAITING states and use

cat jstack.log | grep "java.lang.Thread.State" | sort -nr | uniq -c

to spot problematic threads.

Disk : Check filesystem space with df -hl and monitor I/O performance using iostat -d -k -x. Identify the busiest disks via the %util column and pinpoint the responsible process with iotop. Convert a thread ID to a PID using readlink -f /proc/*/task/<tid>/../.. and inspect its I/O via cat /proc/<pid>/io or lsof -p <pid>.

Memory : Start with free to view overall usage, then differentiate between heap OOM ( java.lang.OutOfMemoryError: Java heap space), native thread stack OOM, and metaspace OOM. Use jmap -histo:live <pid> and Eclipse MAT ( jmap -dump:format=b,file=heap.hprof <pid>) to locate leaks. For off‑heap leaks, monitor native memory with pmap -x <pid> | sort -rn -k3 | head -30, capture dumps via

gdb --batch --pid <pid> -ex "dump memory dump.bin 0x<addr> 0x<addr+size>"

, and analyze with jcmd <pid> VM.native_memory. Enable -XX:+HeapDumpOnOutOfMemoryError for automatic dumps.

GC : Enable detailed GC logging with

-verbose:gc -XX:+PrintGCDetails -XX:+PrintGCDateStamps -XX:+PrintGCTimeStamps

. Use jstat -gc <pid> 1000 to observe Young vs. Full GC frequency. For G1, adjust -Xmn, -XX:SurvivorRatio, -XX:G1ReservePercent, and -XX:InitiatingHeapOccupancyPercent to mitigate frequent or long‑running collections. Dump heap before/after Full GC with jinfo -flag +HeapDumpBeforeFullGC <pid> and jinfo -flag +HeapDumpAfterFullGC <pid>.

Network : Diagnose timeouts, TCP queue overflows, and RST packets. Use netstat -s | egrep "listen|LISTEN" to view overflow counters, ss -lnt to check backlog sizes, and adjust kernel parameters ( net.ipv4.tcp_tw_reuse=1, net.ipv4.tcp_tw_recycle=1) to recycle TIME_WAIT sockets. Monitor connection states with

netstat -n | awk '/^tcp/ {++S[$NF]} END {for(a in S) print a, S[a]}'

or ss -ant. Capture RST traffic with tcpdump -i en0 tcp -w capture.cap and analyze in Wireshark.

By following these systematic checks and leveraging the listed commands, developers can quickly pinpoint the root cause of performance degradations and restore service stability.