async-profiler
github.com
- Free plan
- macOS
- Linux
- API
What async-profiler is
async-profiler is a free, self-hosted sampling profiler for Java that measures performance while avoiding safepoint bias. It works with OpenJDK and other runtimes based on the HotSpot JVM, and its stack traces can show non-Java threads alongside native and kernel frames. Profiling targets include CPU time, Java-heap and native-memory allocations, memory leaks, contended locks, and hardware or software performance counters. Documented modes cover CPU, allocation, wall-clock, Java-method, and multiple-event profiling. Results can be saved as collapsed stacks, flame graphs, HTML tree views, or text, including an interactive Flame Graph for browser viewing. Official builds are maintained for Linux and macOS on x64 and arm64. Developers can use the Java API as a Maven dependency, and IntelliJ IDEA includes async-profiler in its Java Profiler settings. It can profile Java processes in Docker or LXC from within a container or from the host. Building from source requires make, GCC 7.5.0+ or Clang 7.0.0+, static libstdc++, and JDK 11+.
Who it suits
It suits Java developers who need to inspect CPU, memory, locks, or performance counters on HotSpot-based runtimes. Teams profiling containerized Java processes or working from IntelliJ IDEA may also find its integrations useful.
Good to know, and what to watch
- Profiles CPU, heap, native memory, leaks, and locks
- Shows native and kernel frames in stack traces
- Saves results in several formats, including browser-viewable flame graphs
- Java API is available as a Maven dependency
- Can profile Java processes in Docker or LXC
- Official builds are listed only for Linux and macOS
- Source builds require JDK 11 or newer
- Source builds require GCC 7.5.0+ or Clang 7.0.0+
Plans
Every plan and priceCommunity open-source
Free plan
The verdict
async-profiler covers a broad range of Java and native performance signals, with multiple output formats and container support. Building it yourself requires the listed compiler and JDK prerequisites.
What the maker publishes
- Build requirements
- Building requires make, GCC 7.5.0+ or Clang 7.0.0+, static libstdc++, and JDK 11+. source
- Container profiling
- It can profile Java processes in Docker or LXC containers from inside the container or from the host system. source
- IDE integration
- IntelliJ IDEA bundles async-profiler and exposes it through the Java Profiler settings. source
- Java integration
- A Java API is published to Maven Central and can be used as a Maven dependency. source
- Native visibility
- It monitors non-Java threads and shows native and kernel frames in stack traces. source
- Nightly retention
- Nightly binaries for previous commits are kept for 30 days. source
- Official builds
- Officially maintained builds are provided for Linux x64 and arm64 and macOS x64 and arm64. source
- Output
- The profiler can save results as an interactive Flame Graph viewable in a browser. source
- Output formats
- Supported output formats include collapsed stacks, flame graphs, HTML tree views, and text. source
- Profiling modes
- The documentation lists CPU, allocation, wall-clock, Java-method, and multiple-events profiling modes. source
- Profiling targets
- It can profile CPU time, Java-heap allocations, native-memory allocations and leaks, contended locks, and hardware or software performance counters. source
- Purpose
- async-profiler is a low-overhead sampling profiler for Java that avoids the safepoint bias problem. source
- Runtime support
- It works with OpenJDK and other Java runtimes based on the HotSpot JVM. source
- Security reporting
- Security issues should be reported to AWS/Amazon Security through its vulnerability reporting page instead of a public GitHub issue. source
- Span API
- The Span API records latency-sensitive intervals into the same JFR recording as profiling samples and is allocation-free. source
Checked against the maker's pages on 1 Oct 2026.