Please subscribe to our YouTube channel @ https://www.youtube.com/@DevoxxForever
Subscribe to LinkedIn @ https://www.linkedin.com/company/voxxed-days-amsterdam
Follow us on Twitter @ https://twitter.com/voxxedamsterdam

Painful days wasted optimizing the wrong part of the code... Blind attempts to fix a production performance issue. Then, speculative “micro-optimizations" that clutter the code for no measurable gain. Only to later trace the problem to the most unexpected cause. This talk explains how to efficiently zoom in on a performance bottleneck using OpenTelemetry distributed tracing, metrics, and execution profiling. After we learn to read a flame graph, we’ll trace typical backend bottlenecks, such as resource starvation (threads and JDBC Connections), time-consuming interceptors, repeated network calls, CPU hotspots, and lock contention, using a variety of mainstream tools. Join this session to prepare for the performance issues that any successful application will eventually have.

Devoxx

A conference talk transcript covering Java performance bottleneck investigation techniques. Demonstrates live profiling using IntelliJ profiler, Java Flight Recorder (JFR), and GlowRoot to identify real-world issues: an O(n²) HashSet.removeAll bug, a synchronized block holding a DB connection during network calls, unnecessary @Transactional on read methods causing connection pool exhaustion, and lazy-loading triggered by toString in log statements. Also covers distributed tracing with OpenTelemetry, trace ID propagation across microservices, tail sampling strategies, and using AI agents with MCP to analyze heap dumps and Grafana traces. Concludes with a recommended order: reproduce → trace → metrics → profiling.

Tracing Java Performance Bottlenecks by Victor Rentea