In manufacturing systems, a state of high resilience is always desirable. However, internal and external complexity has great influence on these systems. An approach is to increase manufacturing robustness and responsiveness—and thus resilience—by manufacturing control. In order to execute an effective control method, it is necessary to provide sufficient information of high value in terms of data format, quality and time of availability. Nowadays, raw data is available in large quantities. An obstacle to manufacturing control is the short-term handling of events induced by customers and suppliers. These events cause different kinds of turbulence in manufacturing systems. If such turbulences could be evaluated in advance, based on data processing, they could serve as aggregated input data for a control system. This paper presents an approach how to combine turbulence evaluation and the derivation of measures into a learning system for turbulence mitigation. Integrated in manufacturing control, turbulence mitigation increases manufacturing resilience and strengthens the supply network’s resilience.