Introduction: The alternative lengthening of telomeres (ALT) mechanism was first observed in the model organism S. cerevisiae. Interestingly, this mechanism is necessary for the viability of some tumor cells. Unfortunately, its molecular underpinnings are not yet completely understood. Objective: Here, we combine carefully designed non-targeted mass spectrometry-based metabolomics experiments with a bioinformatics approach to characterize the ALT positive phenotype observed in yeast at the metabolomics level. Methods: We profiled the metabolome using mass spectrometry in yeast strains that have lost telomerase expression, as well as that in pre-senescence and the rescued states. To dissect unwanted technical variation from biologically relevant variation between these states, we used a two-step normalization strategy, i.e., first, an empirical Bayesian framework; and next, we corrected for second-order technical effects. Results: Our results show that ALT-positive yeast strains present two different types of metabolic responses to the genetically-induced telomerase dysfunction: (i) systemic and (ii) specific. The key-difference between these responses is that the systemic response lasts even after the yeast strains have been genetically rescued, while the specific response does not. Interestingly, these metabolic changes can be associated with generic stress responses (e.g., DNA damage) as well as specific responses like accelerated aging of early telomerase-inactivation. Conclusions: A mass spectrometry-based metabolomics approach reveals two distinct types of metabolomics response to telomerase dysfunction in yeast. By identifying these changes in protein (e.g., ARG7, and ARG1), and metabolite (e.g., dATP, and dDTP) amounts, we complement the available information on ALT at the genome-wide level.