ATLAS Trigger Monitoring and Operation in Proton Proton Collisions at 900 GeV

The trigger of the ATLAS-experiment is build as a three level system. The first level is realized in hardware while the higher levels (HLT) are pure software implemented triggers based on large PC farms. According to the LHC bunch crossing frequency of 40 MHz and the expectation of up to 23 interactions per bunch crossing at design luminosity, the trigger system must be able to deal with an input rate of 1 GHz whereas the maximum storage rate is 200 Hz. This complex data acquisition and trigger system requires a reliable and redundant diagnostic and monitoring system. This is inevitable for a successful commissioning and stable running of the whole experiment. The main aspects of trigger monitoring are the rate measurements at each step of the trigger decision at each level, the determination of the quality of the physics objects candidates to be selected at trigger level (as candidates for electrons, muons, taus, gammas, jets, b-jets and missing energy) and the supervision of the system's behavior during the data taking. This information has to be presented to shift persons and trigger experts in a convenient way. Additionally, a well understood data quality of the trigger is essential for all physics analyzes. Therefore, the trigger diagnostic tools are fully mirrored to the offline side and the determination of the quality of the trigger objects is part of the standard event reconstruction. With the first proton proton colli sions at in December 2009, the full chain of the ATLAS trigger and data acquisition could be tested with real conditions for the first time. The functionality of the ATLAS trigger diagnostic and monitoring tools is described based on the experience of the successful data taking whit the first collision data covering both, the on- and offline aspects.