Artificial intelligence is having an increasing impact on society and the economy. The associated algorithms are confronted with the challenge of operating with unknown data and being susceptible to changes in their environment. Therefore, post-deployment monitoring is a fundamental aspect of quality assurance that is becoming increasingly important. However, post-deployment monitoring of these algorithms and the systems on which they operate is still at a very early stage. It is therefore still unclear how certain non-functional requirements can be effectively monitored. In this master’s thesis, a systematic literature review is used to create a taxonomy that includes metrics, techniques and influencing factors for monitoring the non-functional requirements of performance, robustness and fairness.