UI overlap, a phenomenon where one UI component visually covers another, is a core feature in modern UI design. This overlap is both a necessary means to construct rich visual hierarchies and interactive experiences, and a root cause of usability issues and performance bottlenecks, such as overdraw. Existing studies proposed methods to detect specific performance-related UI overlap issues. However, there is currently a lack of systematic, data-driven understanding of the state of UI overlap.

To bridge this gap, we conduct the first large-scale empirical study on UI overlap in the \OH ecosystem. As a new mobile system, the user interfaces in \OH applications evolve rapidly and have complex structures. We analyze 100 popular apps, classifying 33,262,624 overlap instances through a novel three-tiered taxonomy. Our findings reveal that \textit{high-cost occlusion} is a critical and previously hard to detect performance defect. This scenario occurs when resource-intensive components are rendered while visually obscured. We propose \tool, an innovative tool that leverages multimodal vision-language models (VLMs) to automatically detect such issues. We construct \bench, a new benchmark for this problem, to evaluate \tool. Our experiments show that \tool successfully identifies 34 out of 38 high-cost occlusion cases in the benchmark, achieving a precision of 97.14% and an F1-score of 93.12%. Our study not only provides a first comprehensive understanding of UI overlap in \OH but also offers a practical tool to automatically diagnose complex performance-related UI bugs. Our benchmark and tools are publicly \href{https://github.com/SMAT-Lab/HapOverlap}{available}.