Concept Drift’s issue is a decisive problem of online machine learning, which causes massive performance degradation in the analysis. The Concept Drift is observed when data’s statistical properties vary at a different time step and deteriorate the trained model’s accuracy and make them ineffective. However, online machine learning has significant importance to fulfill the demands of the current computing revolution. Moreover, it is essential to understand the existing Concept Drift handling techniques to determine their associated pitfalls and propose robust solutions. This study attempts to summarize and clarify the empirical pieces of evidence of the Concept Drift issue and assess its applicability to meet the current computing revolution. Also, this study provides a few possible research directions and practical implications of Concept Drift handling.