Java teams can now see which methods use the most CPU and memory in production, without reproducing the problem locally. Atatus Continuous Profiling for Java samples your running JVM with low overhead, so it can stay on all the time.
Read the hottest code paths in an interactive flame graph, compare two time ranges to see what got slower after a deploy, and let Insights flag memory leaks and sudden spikes for you.
What Java profiling shows you
- CPU flame graphs: See CPU time across the whole call stack, with the most expensive methods drawn widest. Click a frame to zoom in.
- Memory allocation profiling: Find the methods that allocate the most memory to track down leaks and garbage collection pressure before they end in an OutOfMemoryError.
- Lock and wall-clock profiles: Measure lock contention and wall-clock time to find threads that spend their time waiting instead of working.
- Compare profiles: Put a baseline next to a comparison, such as before and after a deploy, and see which methods grew.
- Insights: Memory leaks, sudden spikes, resource exhaustion and long-running CPU trends are flagged for you, with the service and profile type for each.
- Simple setup: Start the JVM with the Atatus profiler agent and set your API key and application name. See the Java profiling docs.