Druv Ghimire@Druv_Ghimire

Modern applications can experience performance degradation when traffic, workload, or resource demands increase unexpectedly. Engineering teams often discover these limitations only after deployment, while repeatedly stress-testing real systems can be expensive, time-consuming, and risky. TRACR addresses this problem through an Application Digital Twin: a simplified, data-driven virtual representation of an application's behavior under different workload conditions. TRACR processes application and workload data such as request volume, response time, throughput, CPU and memory utilization, and error rates. This data is transformed into an analytical model that captures relationships between workload intensity and application performance. Users interact with TRACR through an interactive interface where they can modify workload parameters and explore hypothetical scenarios. The system simulates how the application could respond as traffic or resource demands change, while visualizations expose performance trends, thresholds, and potential bottlenecks. The primary target users are developers, data scientists, DevOps/MLOps teams, and engineering teams involved in application development, testing, and capacity planning. TRACR's key differentiator is that it goes beyond conventional monitoring. Instead of only showing what is happening to an application at the present moment, it provides a "what-if" environment where users can experiment with different workload conditions without directly stressing the real application. The project combines data engineering, analytics, machine learning, simulation, and interactive visualization into a single workflow. Raw telemetry is converted into structured information, analyzed to understand application behavior, and then used to create an interactive representation that users can experiment with. By providing a virtual environment for exploring application behavior, TRACR can help teams identify potential performance limitations earlier, understand scalability characteristics, and make more informed decisions about application capacity and optimization.

💻 Online Hackathon | ⏳ Build: 48 hours | 📅 September 25–27, 2026 The official IBM Bob 2.0 hackathon. Build what's next in AI-assisted development, alongside the developers and experts shaping it. ⭐ Get help from experts throughout the event 🤝 Join a team or build solo 💰 Prize pool: $12,000 🧑🏻‍💻 Register before the Kick-Off Stream to start building with Bob