This dataset was collected during the experimental evaluation of the anomaly-aware monitoring method and has been published to support the related publication titled On Anomaly-Aware Structural Health Monitoring at the Extreme Edge authored by David Arnaiz (corresponding author david.arnaiz@upc.edu), Eduard Alarcón, Francesc Moll, and Xavier Vilajosana. The benchmark structure is a steel frame reduced-scale model structure, built and maintained by the Department of Civil and Environmental Engineering at the Polytechnic University of Catalonia (UPC). The dataset was obtained using two sensor nodes placed on the structure, and contains 275 different monitoring events. The dataset contains the 3-D acceleration data measured by the nodes, the feature vector computed by the nodes, and (for the test cases) the inference results from their anomaly detection models. More information about the dataset can be seen in the README file and the corresponding publication. This work has been made possible thanks to the funding of the Agència de Gestió d’Ajuts Universitaris de Recerca (AGAUR), grand number 2019 DI 075.