FABRIC introduces programmable data planes with P4-enabled switches, FPGAs, and smartNICs. However, debugging large-scale network experiments becomes increasingly complex since network-level visibility alone does not provide sufficient insight into stateful data plane processing. We present and demonstrate a working prototype of ProvP4, a provenance-informed system built on CREASE that captures data plane process dependencies and that enables investigative analysis of reconfigurable data-plane hardware and reproducibility for FABRIC experiments. Our goal is to allow FABRIC users to incorporate stateful data-plane elements into the diagnosis of network issues such as misconfigurations.