Profiling Java apps: breaking things to prove it works

Coroot already does eBPF-based CPU profiling for Java. It catches CPU hotspots well, but that's all it can do. Every time we looked at a GC pressure issue or a latency spike caused by lock contention, we could see something was wrong but not what. We wanted memory allocation and lock contention profiling. So we decided to add async-profiler support to coroot-node-agent. The goal: memory allocation and lock contention profiles for any HotSpot JVM, with zero code changes. Here's how we got there. Why async-profiler We went with async-profiler. It's a native JVMTI agent used by pretty much everyone in the Java profiling space (Pyroscope, IntelliJ, Datadog). It can be loaded into a running JVM dynamically, supports CPU, allocation, and lock contention profiling in a single session, and works in unprivileged containers with no JVM flags. It outputs JFR format, which we parse using Grafana's jfr-parser. How we integrated it The integration follows the same pattern as our Java TLS agent: The