This contribution focuses on performance and operating conditions of the ATLAS Pixel detector with special emphasis on radiation damage. It discusses charge collection properties in data and their modelling with radiation damage simulation, 3D sensor performance with comparison to planar sensors as a function of fluence and the efficiency for associating pixel hits to reconstructed charged particle tracks in relation to operation at high rate and mitigation techniques adopted for LHC Run 3.

At particle fluence of O(10$^{15}$) neutron-equivalent cm$^{-2}$ on the layers closest to the LHC beam pipe, radiation damage has become relevant to most aspects of detector and related physics object performance. ATLAS simulation includes by default a radiation damage pixel digitiser accounting for detailed radiation effects in the Si bulk. The evolution of charge collection efficiency for the innermost barrel pixel layers (IBL and B-Layer) with integrated luminosity and particle fluence has been studied on data and simulation showing good agreement and practicable projections until the end of Run 3. The study of charge collection efficiency for IBL sensors of planar and 3D design confirms the improved response of 3D pixels after irradiation with a gain of $\simeq$ 25-30% in charge collection efficiency at a fluence of $10^{15}$ n-eq cm$^{-2}$ compared to planar sensors. The efficiency for pixel hits-on-track remains constant with time, thanks to the periodic increase of the operating bias voltage, the optimisation of analog thresholds on innermost layers and the mitigation of de-synchronisation effects at high rate with improved DAQ firmware and software.

The ATLAS Pixel detector performs under conditions that are, for some key parameters (pile-up of 50-60 with tests up to 70 and particle fluence from current 1.0-1.3 $\times$ $10^{15}$ to 1.5-2.2 $\times$ $10^{15}$ n-eq cm$^{-2}$ at the end of Run 3) within factors of $\sim$ 3 to 5 compared to the goals for the HL-LHC tracker upgrades. These results already provide useful indications for the optimisation of the operating conditions for the new generation of pixel trackers.