An application that performed well yesterday may become slower today. But identifying what changed and determining the reason behind the degradation can be challenging. Was it increased garbage collection activity, slower method execution, excessive memory allocation, thread contention, network delays, or a surge in exceptions?
Traditional monitoring helps detect performance degradation, but identifying the underlying changes often requires engineers to investigate multiple metrics and diagnostic data sources.
In this webinar, Ram Lakshmanan, Architect at yCrash, shared how comparative diagnostics and AI-assisted analysis can help engineering teams identify performance regressions by comparing more than 200 performance signals across two application runs. With his expertise in Java application performance, JVM internals, and production troubleshooting, Ram demonstrated how engineers can distinguish meaningful performance changes from normal variations, investigate potential root causes, and validate performance optimizations.
Why Detecting Performance Regressions Matters
Performance regressions can occur after application deployments, configuration changes, code modifications, or infrastructure updates. Even when individual monitoring metrics appear normal, changes across multiple performance indicators can reveal underlying issues.
For engineering teams, identifying these changes requires more than looking at isolated metrics. They need to understand how application behavior has changed between two runs and determine which differences are significant.
Comparative performance analysis makes this possible by examining multiple diagnostic signals together. With AI-assisted analysis, engineers can interpret these differences more efficiently and narrow down potential causes of performance degradation.
Key Takeaways
- Compare 200+ performance signals across two application runs to identify regressions and improvements.
- Identify changes in GC activity, method execution, memory allocation, thread contention, network I/O, and exceptions.
- Distinguish meaningful performance changes from normal variations and diagnostic noise.
- Use comparative diagnostics to narrow down potential root causes of performance regressions.
- Apply AI-assisted analysis to interpret performance differences and accelerate troubleshooting.
- Validate performance optimizations by comparing application behavior before and after changes.
- Understand how comparative diagnostics can improve application performance troubleshooting.
Slide Deck
You can now access the presentation slides used during the webinar to revisit the concepts, performance signals, and comparative diagnostic approaches discussed during the session.
Webinar Recording
Watch the complete webinar recording to understand how comparative diagnostics across 200+ performance signals can help identify application performance regressions.
Stay Tuned
Identifying performance regressions early is essential to maintaining application reliability and performance.
Stay connected with yCrash for upcoming webinars featuring practical insights into Java performance engineering, JVM troubleshooting, production diagnostics, and AI-assisted application performance analysis.
📌 Click here to explore our upcoming webinars.
Don’t miss our next webinar!