Frozen threads

• Goals of frozen threads
• Frozen thread visibility
• Toolbar

Goals of frozen threads

Frozen threads view in the profiler helps identify threads that remain in the same state for an extended period, which may indicate potential deadlocks or hung threads. This capability complements the standard deadlocks detector by uncovering issues that might otherwise go unnoticed, such as deadlocks caused by JVM internal locks or native methods using low-level synchronization primitives.

Frozen thread detection depends on thread profiling with stack recording enabled. If thread profiling is stopped, the profiler will not be able to detect frozen threads. By continuously monitoring thread activity and stack changes, the profiler accurately identifies threads that appear stuck or unresponsive over time.

Frozen thread visibility

By default, well-known cases where threads remain in the same state, such as those waiting in a socket or JVM internal thread, are excluded from view to highlight potentially problematic threads.

You can customize visibility using and buttons, filtering out threads performing long but legitimate operations or revealing hidden threads for further evaluation. This ensures you can focus on the most relevant threads during analysis.

Toolbar

• : Show all frozen threads, including hidden ones.
• : Apply filters to this view.
• : Expand all threads to show stacks.
• : Collapse all threads to hide stacks.
• : Copy all stack traces.