Scalable Tracy Integration for HPX: Unified Task Lifecycle and Causal Tracing

HPX already has partial Tracy support and stronger integrations with tools like VTune and APEX, but tracing is still fragmented and largely thread-centric. This makes it difficult to understand how tasks actually behave, especially across suspension, resumption, and asynchronous dependencies. This project focuses on strengthening and extending Tracy integration in HPX by building on the existing hpx::tracing abstraction and expanding task-level instrumentation across key runtime paths. The goal is to make task execution easier to follow by improving lifecycle visibility, especially around suspend/resume behavior, and by adding initial causal signals (for example, future fulfillment and other synchronization events) to help explain why tasks resume. The work will also extend tracing to execution paths that currently lack visibility, such as fork_join_executor and continuation execution, and introduce scalable controls like filtering and sampling to keep tracing usable in fine-grained workloads. As a stretch goal, I will explore optional trace-context propagation across parcel paths to enable cross-locality correlation in distributed runs. The expected deliverables include integrated tracing hooks across core runtime components, improved suspend/resume tracing, better coverage of executor and continuation paths, benchmark-validated low-overhead operation, and clear documentation for both maintainers and users.