Automating CMS Calibrations using the WMAgent framework

Particle beams are now circulating in the world's most powerful particle accelerator LHC at CERN and the experiments are ready to record data from beam. Data from first collisions will be crucial for sub-detector commissioning, making alignment and calibration high priority activities. Executing the alignment and calibration workflows represent a complex and time consuming task, with intricate data dependencies and stringent latency requirements. For this reason CMS Workload Management has defined an architecture and strategy to automate the workflows to minimise the required human effort. In this paper, we discuss the WMCore and WMAgent components used in the current CRABSERVER framework showing the prototype tuned over the first use case, namely calibration of the Beam Spot. We also present the longer term strategies for moving alignment and calibration workflows into the standard Tier-0 processing and automating the analysis workflow.