Assessment of data quality in ATLAS

Abstract

Assessing the quality of data recorded with the ATLAS detector is crucial for commissioning and operating the detector to achieve sound physics measurements. In particular, the fast assessment of complex quantities obtained during event reconstruction and the ability to easily track them over time are especially important given the large data throughput and the distributed nature of the analysis environment. The data are processed once on a computer farm comprising O(1, 000) nodes before being distributed on the Grid, and reliable, centralized methods must be used to organize, merge, present, and archive data-quality metrics for performance experts and analysts. A review of the tools and approaches employed by the detector and physics groups in this environment and a summary of their performances during commissioning are presented.