Trigger Monitoring at ATLAS

The Trigger and Data Acquisition system for the ATLAS experiment has to reduce the 40 MHz of LHC bunch crossing rate to ~200 Hz of recording rate. This is achieved through a complex distributed system composed by $sim$ 1.000 CPUs, about a third of the expected final size of the system. Monitoring the trigger behavior through all the trigger level is of fundamental importance to assess the quality of the data taken, to give fast feedback for the trigger configuration design and to monitor the stability of the HLT farm components. In this paper we will present the online monitoring framework and the various tools available in the ATLAS trigger system going from the ones that build the basic monitoring infrastructure and test the basic functionalities of the system to the more elaborated ones that checks the quality of the data taking looking at physics variables reconstructed online. The early experience in the 2009 cosmics data taking period will also be shown.