Graph machine learning (GML) has proliferated in domains closely entwined with human lives, thereby amplifying the imperatives of fairness within these computational models. While machine learning fairness has witnessed significant progress, setting foundational guidelines for its graphbased counterpart, several intricacies exclusive to graph structures necessitate dedicated solutions. This survey delineates recent strides in graph ML fairness, encapsulating methodologies to both comprehend and augment fairness in graph ML systems. We illuminate the constraints of prevailing studies through rigorous formulaic analysis and systematically categorize them by their fairness criteria, debiasing strategies, and empirical models and tasks. Furthermore, we prognosticate forthcoming challenges, stemming not only from overarching ML fairness concerns but also from the inherent characteristics unique to graphs.