Which three JVM options would you add to investigate performance issues in a HotSpot JVM?

Question

Exhibit

Options

• Axx:+PrintGCDetails
• Bxx:+PrintGCTimeStamps
• Cxx:+HandlePromotionFailure
• Dverbose:gc
• Eserver

• A
  86% (31)
• C
  8% (3)
• E
  6% (2)

Explanation

JVM Performance Investigation Options

Correct Answers: A, B, D

Note: The options appear to be missing the leading - dash - in practice these are -XX:+PrintGCDetails, -XX:+PrintGCTimeStamps, and -verbose:gc.

Option-by-Option Breakdown

A: -XX:+PrintGCDetails YES

Enables verbose GC logging with full detail - heap sizes before/after collection, pause duration, and which GC regions were affected. This is a primary diagnostic tool for spotting memory pressure and long pause times.

B: -XX:+PrintGCTimeStamps YES

Adds timestamps to GC log entries so you can correlate GC events with application slowdowns on a timeline. Alone it's minimal, but combined with PrintGCDetails it becomes powerful for pattern analysis.

C: -XX:+HandlePromotionFailure NO

This is a GC tuning/configuration flag, not a diagnostic one. It allowed minor GC to proceed even when the old generation lacked space for all promoted objects. It was deprecated and removed in Java 6u24+ - it doesn't investigate anything, it changes GC behavior.

D: -verbose:gc YES

A standard JVM flag (not an -XX extension) that enables basic GC output - when collections occur and how long they take. It's the simplest GC diagnostic available and works across all JVM implementations.

E: -server NO

This selects the server-mode JIT compiler (C2), which applies more aggressive optimizations. It changes how the JVM runs code - it's a performance configuration choice, not a diagnostic tool for investigating existing issues.

Memory Tip

"Investigate = observe, not change"

If a flag modifies JVM behavior, it's tuning - not investigation. If it outputs diagnostic data, it's investigation.

No community discussion yet for this question.